System

Abstract

To provide a system.SOLUTION: A system includes means for receiving inquiry contents from a terminal, means for analyzing the received inquiry contents and filtering out specific wording, means for generating a prompt sentence based on the filtered inquiry contents, means for sending the generated prompt sentence to a generative AI model, means for receiving a response from the generative AI model, means for analyzing the received response and converting it into a format that is easy for a user to understand, and means for sending the converted response to the terminal.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including a description and related instruction sentence regarding the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2022-180282 A Summary of the Invention [Problem to be solved by the invention]

[0004] Some companies want to use generative AI for business purposes, but are concerned about security. In particular, there is a demand for security measures when making inquiries to generative AI, as well as governance such as blocking certain text. [Means for solving the problem]

[0005] We have developed a system that provides a secure proxy SaaS service for generative AI. This system installs a server within the company and queries the generative AI via the server. Server management utilizes operational and security knowledge cultivated in the information and communications industry, enabling governance such as blocking specific wording through settings for each individual company. This makes it possible to provide a secure generative AI environment. [Brief description of the drawings]

[0006] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Diagram 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. FIG. [Diagram 3] FIG. 11 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Diagram 5] FIG. 13 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 13 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 13 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 2 is a sequence diagram showing a process flow of the data processing system in the first embodiment of the first form example. [Figure 12] FIG. 11 is a sequence diagram showing a process flow of the data processing system in the application example 1 of the first embodiment. [Figure 13] FIG. 11 is a sequence diagram showing a process flow of the data processing system in the second embodiment of the second embodiment. [Figure 14] FIG. 13 is a sequence diagram showing a process flow of the data processing system in application example 2 of embodiment 2. [Figure 15] FIG. 13 is a sequence diagram showing a process flow of a data processing system in a third embodiment of the present invention. [Figure 16] FIG. 13 is a sequence diagram showing a process flow of the data processing system in the third application example of the third embodiment. [Figure 17] FIG. 13 is a sequence diagram showing the flow of processing in the data processing system in the first embodiment of the first form example when an emotion engine is combined. [Figure 18] FIG. 13 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Form Example 1 when combined with an emotion engine. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0007] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0008] First, the terms used in the following description will be explained.

[0009] In the following embodiments, a signed processor (hereinafter simply referred to as a "processor") may be one arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be one type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), or a TPU (TENSOR PROCESSING UNIT (registered trademark)).

[0010] In the following embodiments, a signed RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by the processor.

[0011] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0012] In the following embodiments, a communication I / F (Interface) with a code is an interface including a communication processor and an antenna. The communication I / F controls communication between multiple computers. An example of a communication standard applied to the communication I / F is a wireless communication standard including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. In addition, in this specification, the same idea as "A and / or B" is also applied when three or more things are expressed by connecting them with "and / or."

[0014] [First embodiment]

[0015] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0016] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0017] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a wide area network (WAN) and / or a local area network (LAN).

[0018] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0019] The reception device 38 includes a touch panel 38A and a microphone 38B, and receives user input. The touch panel 38A detects contact with a pointer (e.g., a pen or a finger) to receive user input by the touch of the pointer. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0020] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (e.g., voice and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs voice according to instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a Complementary Metal-Oxide-Semiconductor (CMOS) image sensor or a Charge Coupled Device (CCD) image sensor.

[0021] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54.

[0022] FIG. 2 shows an example of main functions of the data processing device 12 and the smart device 14.

[0023] As shown in Fig. 2, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32. The specific process program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific process program 56 from the storage 32, and executes the read specific process program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific process program 56 executed on the RAM 30.

[0024] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0025] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores a reception output program 60. The reception output program 60 is used together with the specific processing program 56 by the data processing system 10. The processor 46 reads out the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0026] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0027] "Example 1"

[0028] The present invention is a system that provides a secure proxy SaaS service for generative AI. Specifically, a server is installed within the company, and queries are made to the generative AI via this server. Server management utilizes operational and security knowledge cultivated in the information and communications industry, enabling governance such as blocking specific wording through settings for each individual company.

[0029] "Example 2"

[0030] As a concrete example, if a company wants to use generative AI to respond to customers, they can use the system of the present invention to make inquiries to the generative AI via a server installed within the company. In this case, server management can utilize the operational and security knowledge cultivated in information and communications to set up settings to block specific text, such as customer information or confidential information. This makes it possible to use generative AI in business while ensuring security.

[0031] The process flow of each embodiment will be described below.

[0032] "Example 1"

[0033] Step 1: Set up a server within the company. This server acts as an intermediary for communication with the generative AI.

[0034] Step 2: A query is sent to the generative AI via the server. This query is based on the company's business needs, for example, generating an automatic response for customer support.

[0035] Step 3: Manage the server. This management utilizes operational and security knowledge cultivated in the information and communications industry, enabling governance such as blocking specific wording with settings for each individual company.

[0036] "Example 2"

[0037] Step 1: If a company wants to use generative AI to handle customer support, it first sets up a server within the company.

[0038] Step 2: Make an inquiry to the generative AI via the installed server. This inquiry may involve generating an automatic response to a customer inquiry.

[0039] Step 3: Manage the server. This management utilizes operational and security knowledge cultivated in the information and communications industry, and sets up settings to block specific text, such as customer information or confidential information. This makes it possible to use generative AI in business while ensuring security.

[0040] Example 1

[0041] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal".

[0042] In order to use generative AI models safely and efficiently within a company, it is necessary to minimize the risk of data leakage to the outside while implementing governance based on the company's security policy. However, conventional systems lack the ability to block or filter specific text, making it difficult to fully comply with a company's security policy.

[0043] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0044] In this invention, the server includes a means for providing a secure proxy service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving a prompt text from a user's terminal, a means for inspecting the received prompt text based on filtering rules, a means for relaying the prompt text that has passed the inspection to the generative AI model, a means for receiving a response from the generative AI model and inspecting it again based on filtering rules, and a means for transmitting the response that has passed the filtering to the user's terminal. This makes it possible to safely and efficiently use the generative AI model while realizing governance based on the company's security policy.

[0045] The "Secure Proxy Service for Generative AI" is a service that securely relays queries to generative AI models and enables governance based on a company's security policies.

[0046] A “server” is a computer system installed within a company that relays queries to the generative AI model and provides filtering and governance functions.

[0047] A "prompt" is text that contains instructions or questions that a user enters into a generative AI model.

[0048] "Filtering rules" are rules or conditions for blocking specific words or phrases, and are set based on a company's security policy.

[0049] A "generative AI model" is an artificial intelligence model that generates responses based on prompts from the user.

[0050] "Governance" refers to management and control, such as blocking specific words or phrases, based on a company's security policy.

[0051] A "user's device" is a device, such as a computer or smartphone, that a user uses to input prompts to a generative AI model and receive responses.

[0052] An "inquiry" refers to instructions or questions that a user gives to a generative AI model, and is sent to the server as a prompt.

[0053] A "response" is an answer or piece of information that a generative AI model generates based on a prompt.

[0054] This invention is a system that provides a secure proxy service for generative AI, and queries generative AI models through a server installed within a company. A specific embodiment of this system will be described below.

[0055] Server installation and initial configuration

[0056] The server is installed within the company and connected to the company's network. The server administrator accesses the server's settings screen and sets filtering rules based on the company's security policy. For example, a rule is set to block inquiries that include the phrase "confidential information." Once the settings are complete, the server is connected to the company's network and begins operation.

[0057] User query input

[0058] The user inputs a prompt for the generative AI model from their own device. For example, a marketing person inputs a prompt such as, "Please tell me your sales forecast for the next quarter." Once the input is complete, the user clicks the send button.

[0059] The server receives and examines the query

[0060] The server receives the prompt text sent from the user's terminal. It checks the received prompt text based on the filtering rules. For example, it checks whether the text contains the phrase "confidential information." If there are no problems, it proceeds to the next step.

[0061] Relaying inquiries

[0062] The server relays the prompts that pass inspection to the generative AI model, which generates a response based on the prompts received. For example, it generates a response like, "Sales forecast for the next quarter is XXX yen."

[0063] Receiving and filtering responses by the server

[0064] The server receives the response from the generative AI model. The server checks the response again based on the filtering rules. For example, it checks whether the response contains the phrase "confidential information." If there are no problems, it proceeds to the next step.

[0065] Providing a response

[0066] The server sends the filtered responses to the user's device. The user can check the response from the generative AI model on the device. For example, the user can check the response, "Sales forecast for the next quarter is XXX yen."

[0067] Examples

[0068] As a concrete example, Company A has a setting to block queries that contain the phrase "confidential information." When a user enters the prompt phrase "Please generate a report that contains confidential information," the server detects this phrase and returns an error message without relaying the query.

[0069] Company B also uses a generative AI model to generate marketing reports. When a user enters a prompt such as "Please generate a report on the latest market trends," the server inspects the query and relays it to the generative AI model if it determines there is no problem. The response from the generative AI model is provided to the user via the server.

[0070] The system enables companies to use generative AI models safely and efficiently.

[0071] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0072] Step 1:

[0073] The server is installed within the company and connected to the network. The server administrator accesses the server's configuration screen and sets filtering rules based on the company's security policy. For example, a rule is set to block inquiries that include the phrase "confidential information." The input is the security policy and filtering rules, and the output is the configured server.

[0074] Step 2:

[0075] The user inputs a prompt sentence to the generative AI model from his / her own device. For example, the user inputs a prompt sentence such as "Please tell me your sales forecast for the next quarter." The input is the prompt sentence, and the output is the sent prompt sentence. The user clicks the send button.

[0076] Step 3:

[0077] The server receives the prompt text sent from the user's terminal. It checks the received prompt text based on filtering rules. For example, it checks whether it contains the phrase "confidential information." The input is the sent prompt text, and the output is the inspection result. If there are no problems, it proceeds to the next step.

[0078] Step 4:

[0079] The server relays the prompts that pass inspection to the generative AI model. The generative AI model generates a response based on the prompts it receives. For example, it generates a response like, "Sales forecast for the next quarter is XXX yen." The input is the prompt that passes inspection, and the output is the generated response.

[0080] Step 5:

[0081] The server receives the response from the generative AI model. The received response is inspected again based on the filtering rules. For example, it checks whether it contains the phrase "confidential information." The input is the response from the generative AI model, and the output is the inspection result. If there are no problems, it proceeds to the next step.

[0082] Step 6:

[0083] The server sends the filtered response to the user's device. The user can check the response from the generative AI model on the device. For example, the user can check the response "Sales forecast for the next quarter is XXX yen." The input is the filtered response, and the output is the response sent to the user's device.

[0084] (Application example 1)

[0085] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0086] When using generative AI within a company, there is a risk that confidential information may be leaked to the outside. In addition, it is difficult to implement appropriate governance when specific wording or data is included. Furthermore, security must be ensured when making inquiries to generative AI using smartphones.

[0087] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means. In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing a server, a means for making an inquiry to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for making an inquiry to the generative AI through an application installed on a smartphone, a means for securely communicating via a secure proxy server, a means for sending a filtered prompt to the generative AI, and a means for displaying the result on a smartphone application. This makes it possible to reduce the risk of confidential information leakage and perform appropriate governance when using generative AI within a company. In addition, high security can be ensured even when making inquiries to the generative AI using a smartphone.

[0088] The "Secure Proxy SaaS Service for Generative AI" is a cloud-based service that provides a proxy server for securely making inquiries to generative AI.

[0089] A "server" is a computer system that relays queries to the generative AI and manages security and governance.

[0090] A "means for querying a generative AI via a server" is a method for sending a prompt to a generative AI via a server and receiving a result.

[0091] "Means of server management" refers to the methods for maintaining and properly managing the operation and security of the server.

[0092] "Measures that enable governance, such as blocking specific wording" refers to methods for filtering out specific keywords and phrases and controlling the content of queries made to generative AI.

[0093] An "application installed on a smartphone" is software that runs on a smartphone and is used to make queries to the generative AI.

[0094] "Means for securely communicating via a secure proxy server" refers to a method for encrypting data and communicating securely through a secure proxy server.

[0095] "Means for transmitting a filtered prompt to the generative AI" refers to a method for transmitting the prompt after it has been filtered to the generative AI.

[0096] "Means for displaying results in a smartphone application" refers to a method for displaying the results returned from the generative AI in an application on a smartphone.

[0097] The following system configuration will be described as an embodiment of the present invention.

[0098] System Configuration

[0099] Hardware

[0100] Server: A computer system installed within a company that provides a secure proxy SaaS service for generative AI.

[0101] Smartphone: A device that allows users to make inquiries to generative AI.

[0102] software

[0103] Secure Proxy Server: Runs on the server, relays queries to the generative AI, and manages security and governance.

[0104] Smartphone application: Software that is installed on a smartphone and allows for queries to be made to the generative AI.

[0105] Data processing and calculation

[0106] server

[0107] The server provides a secure proxy SaaS service for generative AI. When a user makes a query to generative AI through a smartphone application, the prompt is first sent to the server. The server applies governance such as blocking certain words, and sends the filtered prompt to the generative AI.

[0108] Smartphone

[0109] A user uses a smartphone application to make a query to the generative AI. After entering a prompt and pressing the send button, the prompt is sent to the generative AI via a secure proxy server. The results returned from the generative AI are then displayed on the smartphone application via the secure proxy server again.

[0110] Examples

[0111] For example, when a company queries a generative AI for documents containing confidential information, the system can be used to reduce the risk of confidential information leaking to the outside. The user enters the following prompt into a smartphone application:

[0112] Prompt: "Generate the following sentence: Details about a company's new product."

[0113] This prompt is sent to a server where it is filtered to see if it contains specific words, and the filtered prompt is sent to a generative AI, which then displays the generated results on a smartphone application.

[0114] In this way, it is possible to reduce the risk of confidential information leaks and implement appropriate governance when using generative AI within a company. It is also possible to ensure high security when making inquiries to generative AI using smartphones.

[0115] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0116] Step 1:

[0117] The user starts the smartphone application and inputs the question (prompt) they want to ask the generative AI. The input prompt is temporarily saved in the application.

[0118] Step 2:

[0119] When the user presses the send button, the smartphone application sends the prompt text to the secure proxy server. The input is the prompt text, and the output is a send request to the secure proxy server.

[0120] Step 3:

[0121] The server analyzes the received prompt text and filters it to see if it contains specific words. The input is the prompt text, and the output is the filtered prompt text. Specifically, the server checks it against a specific keyword list and removes any inappropriate words that it finds.

[0122] Step 4:

[0123] The server sends the filtered prompt to the generative AI. The input is the filtered prompt, and the output is a request to send to the generative AI. Specifically, the server sends an HTTP request to the API endpoint of the generative AI.

[0124] Step 5:

[0125] The generative AI generates a response based on the received prompt and sends the result back to the server. The input is the filtered prompt and the output is the generated response. In concrete terms, the generative AI uses its internal model to generate a response to the prompt.

[0126] Step 6:

[0127] The server receives the response returned from the generative AI and sends it to the smartphone application. The input is the generated response, and the output is a request to send to the smartphone application. In concrete terms, the server sends an HTTP request with the response data to the API endpoint of the smartphone application.

[0128] Step 7:

[0129] The smartphone application displays the response received from the server. The input is the generated response and the output is the result displayed to the user. In concrete terms, the application passes the received data to a UI component for display on the screen.

[0130] Example 2

[0131] Next, a second embodiment of the second embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0132] Customer support systems that use generative AI are required to respond to customers efficiently while ensuring security. However, with conventional systems, there is a risk that customer information or confidential information may be sent to the generative AI model, which may lead to security issues. In addition, responses from the generative AI model may be difficult for users to understand, which can lead to a decline in the quality of customer support. To solve these issues, a system that can filter the content of inquiries and convert the content of responses is needed.

[0133] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0134] In this invention, the server includes a means for receiving an inquiry from a terminal, a means for analyzing the received inquiry content and filtering out specific wording, a means for generating a prompt sentence based on the filtered inquiry content, a means for sending the generated prompt sentence to a generative AI model, a means for receiving a response from the generative AI model, a means for analyzing the received response and converting it into a format that is easy for a user to understand, a means for sending the converted response to the terminal, a means for installing a server, a means for making an inquiry to the generative AI model via the server, a means for managing the server, and a means for enabling governance such as blocking specific wording. This enables efficient customer service using generative AI while ensuring security.

[0135] A "terminal" is a device through which a user inputs inquiry content and transmits it to a server.

[0136] A "server" is a device that receives queries from a terminal, queries the generative AI model, and sends responses to the terminal.

[0137] "Inquiry content" is text data of a question or request that a user inputs via a terminal.

[0138] "Filtering" is the process of removing or masking specific wording (such as customer information or confidential information) from the content of an inquiry.

[0139] A "prompt sentence" is the text of a query that is generated to be sent to a generative AI model.

[0140] A "generative AI model" is an artificial intelligence model that generates a response based on an input prompt sentence.

[0141] A "response" is the text data of the answer returned from the generative AI model.

[0142] "Analysis" is the process of understanding the inquiry and response, and extracting or converting the necessary information.

[0143] "Transformation" is the process of changing the response from a generative AI model into a form that is easy for a user to understand.

[0144] "Governance" refers to the means of managing the operation and security of a system, and exercising control such as blocking specific text.

[0145] This invention relates to a customer service system that uses generative AI. Specifically, it is a system that uses a terminal, a server, and a generative AI model to filter inquiries and convert responses.

[0146] Hardware and Software Used

[0147] The terminal is a device where a user inputs the contents of an inquiry and transmits the input to the server. The terminal may be a PC, a smartphone, a tablet, or the like.

[0148] The server is a device that receives queries from the terminal, queries the generative AI model, and sends the response to the terminal. The server has installed software for parsing the query content, filtering, generating prompts, and parsing and converting the response. Specific software includes a natural language processing (NLP) library and an API client.

[0149] A generative AI model is an artificial intelligence model that generates a response based on an input prompt sentence. Specifically, GPT-4 (registered trademark) from OpenAI (registered trademark) is used.

[0150] Data processing and calculation

[0151] The terminal transmits the inquiry content input by the user to the server. For example, the user inputs "Please tell me the price of the new product."

[0152] The server analyzes the received inquiry content and filters out specific phrases (such as customer information or confidential information). Based on the filtered inquiry content, it generates a prompt sentence suitable for the generative AI model. For example, it generates a prompt sentence such as "Please tell me the price of the new product."

[0153] The server sends the generated prompt sentence to the generative AI model, which generates a response based on the prompt sentence and sends it back to the server. For example, it generates a response such as "The price of the new product is $100."

[0154] The server parses the response and converts it into a format that is easy for the user to understand. It then sends the converted response to the terminal. The terminal displays the response to the user. For example, the message "The price of the new product is $100" is displayed on the user's screen.

[0155] Examples

[0156] For example, if a company receives an inquiry from a customer saying, "Please tell me the price of a new product," the following process will take place:

[0157] 1. On the terminal, the user enters "Please tell me the price of the new product" and clicks the send button.

[0158] 2. The server analyzes the received data and extracts the text "Please tell me the price of the new product." It then performs a filtering process to check whether the data contains a specific phrase.

[0159] 3. The server generates a prompt sentence based on the filtered text: "Please tell me the price of the new product."

[0160] 4. The server creates and sends an API request to send the generated prompt to GPT-4.

[0161] 5. The server receives the response from GPT-4 and gets the text “The price of the new product is $100.”

[0162] 6. The server parses the received text and replaces technical terms with more common language where necessary.

[0163] 7. The server sends the converted response to the terminal, which displays the received response to the user, and the message "The new product costs $100" appears on the user's screen.

[0164] In this way, businesses can leverage generative AI to deliver efficient customer service while ensuring security.

[0165] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0166] Step 1:

[0167] The terminal transmits the inquiry content entered by the user to the server. As input, the user inputs "Please tell me the price of the new product" and clicks the send button. As output, the inquiry content is transmitted to the server.

[0168] Step 2:

[0169] The server analyzes the inquiry received from the terminal. As input, it receives the text data "Please tell me the price of the new product" sent from the terminal. To process the data, it analyzes the text using natural language processing (NLP) technology and filters out specific phrases (such as customer information or confidential information). As output, it obtains the filtered inquiry content.

[0170] Step 3:

[0171] The server generates a prompt sentence based on the filtered inquiry content. As input, it receives the filtered text "Please tell me the price of the new product." As data processing, it converts it into a format suitable for the generative AI model. As output, it obtains the generated prompt sentence.

[0172] Step 4:

[0173] The server sends the generated prompt sentence to the generative AI model. As input, it receives the generated prompt sentence "Please tell me the price of the new product." As a specific operation, it creates an API request and sends it to the generative AI model (e.g., GPT-4). As output, the prompt sentence is sent to the generative AI model.

[0174] Step 5:

[0175] The server receives the response from the generative AI model. As input, it receives the response data returned from the generative AI model. As a specific operation, it parses the API response and obtains the text "The price of the new product is $100." As output, it obtains the response from the generative AI model.

[0176] Step 6:

[0177] The server analyzes the received response and converts it into a format that is easy for the user to understand. As input, it receives the response data "The price of the new product is $100." As data processing, it performs conversion processing such as replacing technical terms with common terms. As output, it obtains the converted response.

[0178] Step 7:

[0179] The server sends the converted response to the terminal. As input, it receives the converted response data "The price of the new product is $100." As a specific operation, it sends the response data to the terminal. As output, the response is sent to the terminal.

[0180] Step 8:

[0181] The terminal displays the received response to the user. As input, it receives the response data "The price of the new product is $100" sent from the server. As a specific operation, it displays the response message on the user's screen. As output, the user can confirm the response.

[0182] (Application example 2)

[0183] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0184] Conventional customer service systems using generative AI have difficulty in ensuring security and cannot block certain text (e.g., customer information or confidential information), making it difficult to use safely within a company. In addition, they lacked the functionality to make personalized product recommendations to users, making it difficult to improve customer satisfaction.

[0185] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0186] In this invention, the server includes a means for providing a secure proxy service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving questions from users and sending them to the generative AI, a means for returning answers from the generative AI to the user, and a means for making personalized product recommendations based on the user's purchase history and browsing history. This enables safe customer service and personalized product recommendations using generative AI while ensuring security.

[0187] The "Secure Proxy Service for Generative AI" is a relay service that enables safe inquiries to be made to generative AI, and exchanges data while ensuring security.

[0188] A "server" is a computer system that stores, processes, and manages data on a network.

[0189] A "means for querying a generative AI via a server" is a method for sending questions or instructions to a generative AI via a server.

[0190] "Server management" refers to the activity of operating, maintaining, and managing the security of a server.

[0191] "Governance such as blocking specific wording" is a management technique that filters out specific keywords and phrases (such as customer information or confidential information) to prevent the leakage of inappropriate information within the system.

[0192] "Means for receiving questions from a user and sending them to a generative AI" refers to a method for receiving questions entered by a user and sending them to a generative AI.

[0193] "Means of returning an answer from a generative AI to a user" refers to a method for returning an answer obtained from a generative AI to a user.

[0194] "Means for personalized product recommendations based on a user's purchase history and browsing history" refers to a method for analyzing a user's past purchase history and browsing history and recommending individually optimized products based on that.

[0195] A system for implementing this invention includes a server that provides a secure proxy service for the generative AI, a terminal that receives questions from users and sends them to the generative AI, a terminal that returns answers from the generative AI to the user, and a function that makes personalized product recommendations based on the user's purchasing history and browsing history.

[0196] Hardware and Software Configuration

[0197] Server: Installed within a company, it provides a secure proxy service for generative AI. The server is built using a web framework such as Flask.

[0198] Device: A device used by a user, such as a smartphone or tablet. These devices run applications that receive questions from users and send them to the generative AI.

[0199] Generative AI: Uses generative AI models such as OpenAI's GPT-3 (registered trademark). Generative AI generates appropriate answers to questions posed by users.

[0200] Data processing and calculation

[0201] 1. Receiving user questions:

[0202] The user inputs a question through the terminal, for example, "How do you use this product?"

[0203] 2. Filter questions:

[0204] The server filters certain phrases (customer information and confidential information) from the question, ensuring security.

[0205] 3. Querying the generative AI:

[0206] The filtered questions are sent to a generative AI, which generates answers to the questions.

[0207] 4. Response to the answer:

[0208] The server receives the answer from the generative AI and responds to the user through the device, for example, "This product is used in the following steps..."

[0209] 5. Personalized product recommendations:

[0210] The server analyzes the user's purchase and browsing history and recommends individually optimized products based on that information, for example by displaying "These are the products we recommend for you."

[0211] Examples

[0212] When a user asks, "How do you use this product?", the server filters the question and sends it to the generative AI. The generative AI responds with, "This product is used in the following steps..." and the answer is returned to the user. Related products are also recommended based on the user's past purchase and browsing history.

[0213] Examples of prompt statements

[0214] "How do you use this product?"

[0215] In this way, secure customer interactions and personalized product recommendations using generative AI can be achieved while ensuring security.

[0216] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0217] Step 1:

[0218] A user inputs a question through a terminal, for example, "How do you use this product?" The input question is sent to a server by an application on the terminal.

[0219] Step 2:

[0220] The server filters the questions it receives. Specifically, it detects specific phrases (customer information or confidential information) from the questions and blocks or masks them. The filtered questions are then sent to the generative AI.

[0221] Step 3:

[0222] The server sends the filtered questions to the generative AI, which generates appropriate answers to the questions it receives. For example, it uses OpenAI's GPT-3 to generate an answer like, "This product is used in the following steps..."

[0223] Step 4:

[0224] The server receives the answer from the generative AI. The server returns the answer to the device. The device displays the answer from the generative AI to the user.

[0225] Step 5:

[0226] The server analyzes the user's purchase history and browsing history. Specifically, it retrieves the user's past purchase history and browsing history from the database and makes personalized product recommendations based on that. For example, it displays "These are the products we recommend for you."

[0227] Step 6:

[0228] The server sends the personalized product recommendations to the terminal. The terminal displays the recommended products to the user. The user can review the recommended products and consider purchasing them.

[0229] In this way, secure customer interactions and personalized product recommendations using generative AI can be achieved while ensuring security.

[0230] Furthermore, an emotion engine that estimates the emotion of the user may be combined. That is, the identification processing unit 290 may estimate the emotion of the user using the emotion identification model 59, and perform identification processing using the emotion of the user.

[0231] "Example 1"

[0232] As one embodiment of the present invention, there is a system that provides a secure proxy SaaS service for generative AI, which is combined with an emotion engine. This system queries the generative AI through a server installed in the company and manages the server. The server management utilizes operational and security knowledge cultivated in information and communications, enabling governance such as blocking specific wording. Furthermore, the generative AI includes an emotion engine that recognizes the user's emotions. Specifically, it receives text input from the user and recognizes the user's emotions from the text. Then, the generative AI's response is adjusted based on the recognized emotion. For example, if a user inputs text indicating an emotion of anger, the emotion engine recognizes the anger, and the generative AI generates an appropriate response based on the recognition result. This enables a more appropriate response that corresponds to the user's emotions.

[0233] "Example 2"

[0234] As one embodiment of the present invention, there is a system that provides a secure proxy SaaS service for generative AI, which is combined with an emotion engine. This system queries the generative AI through a server installed in the company and manages the server. The server management utilizes operational and security knowledge cultivated in information and communications, enabling governance such as blocking specific wording. Furthermore, the generative AI includes an emotion engine that recognizes the user's emotions. Specifically, it receives text input from the user and recognizes the user's emotions from the text. Then, the generative AI's response is adjusted based on the recognized emotion. For example, if a user inputs text indicating an emotion of anger, the emotion engine recognizes the anger, and the generative AI generates an appropriate response based on the recognition result. This enables a more appropriate response that corresponds to the user's emotions.

[0235] The process flow of each embodiment will be described below.

[0236] "Example 1"

[0237] Step 1: Receive text input from a user, which may include the user's sentiment.

[0238] Step 2: Input the received text into the emotion engine, which recognizes the user's emotion from the text.

[0239] Step 3: The output from the emotion engine (the recognized emotion) is fed into the generative AI, which generates a response based on this emotion.

[0240] Step 4: The generative AI's response is returned to the user. This response corresponds to the user's emotion.

[0241] Example 1

[0242] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal".

[0243] Conventional services using generative AI have security and governance issues, making it difficult to block specific text or operate based on a company's security policy. In addition, they are unable to generate responses that take into account the user's emotions, which can lead to a poor user experience.

[0244] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0245] In this invention, the server includes a means for providing a secure proxy service for the generative AI, a means for installing the server, a means for making an inquiry to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific words, a means for including an emotion engine for recognizing the user's emotions, and a means for adjusting the response of the generative AI based on the emotions recognized by the emotion engine. This makes it possible to generate an appropriate response corresponding to the user's emotions while strengthening security and governance.

[0246] "Generative artificial intelligence" is an artificial intelligence technology that generates natural language responses and content based on user input.

[0247] A "secure proxy service" is a service that relays user requests and provides proxy functions to enhance security and governance.

[0248] A "server" is a computer system that processes and manages data on a network.

[0249] "Governance" refers to the management and control of system operations, such as blocking specific wording and applying security policies.

[0250] An "emotion engine" is software or hardware for analyzing text input from a user and recognizing its emotions.

[0251] "Adjusting the response" means changing the content of the response generated by the generative artificial intelligence based on the emotion recognized by the emotion engine.

[0252] The present invention is a system that provides a secure proxy service for querying a generative artificial intelligence through a server installed within a company. The system includes an emotion engine that recognizes the user's emotions and adjusts responses based on the emotions.

[0253] Hardware and Software Used

[0254] 1. Server

[0255] It is installed within a company.

[0256] It will be managed utilizing operational and security knowledge cultivated in the information and communications industry.

[0257] It has a governance function that blocks specific wording.

[0258] 2. Terminal

[0259] A device (e.g. PC, smartphone) that allows a user to input text.

[0260] 3. Generative Artificial Intelligence

[0261] An artificial intelligence technology that generates natural language responses based on user input.

[0262] 4. Emotion Engine

[0263] Software that analyzes text input from users and recognizes their emotions.

[0264] Data processing and calculation

[0265] 1. User text input

[0266] A user uses the terminal to input text.

[0267] For example: "Why is it so slow?"

[0268] 2. Receiving and parsing the text by the server

[0269] The server receives text input from the user and checks whether it contains specific words.

[0270] Example: Detect the keyword "slow".

[0271] 3. Emotion Recognition by Emotion Engine

[0272] The server sends the text to an emotion engine, which analyzes the text to recognize the user's emotion.

[0273] Example: The emotion engine recognizes "anger."

[0274] 4. Response generation using generative artificial intelligence

[0275] The server sends the emotion recognition results to the generative artificial intelligence, which then generates an appropriate response.

[0276] For example: "We apologize for the inconvenience. We are currently working on improving our system."

[0277] 5. Server Sends Response

[0278] The server sends the generated response to the user's terminal.

[0279] Examples

[0280] For example, if a user enters the text "Why is it so slow?", the following happens:

[0281] 1. User text input

[0282] A user types in the chat window, "Why is this so slow?"

[0283] 2. Receiving and parsing the text by the server

[0284] The server receives the text "Why is it so slow?" and detects the keyword "slow".

[0285] 3. Emotion Recognition by Emotion Engine

[0286] The server sends the text to an emotion engine, which recognizes "anger."

[0287] 4. Response generation using generative artificial intelligence

[0288] The server sends the emotion recognition results to the generative artificial intelligence, which generates a response saying, "We apologize for the inconvenience. We are currently working on improving the system."

[0289] 5. Server Sends Response

[0290] The server displays the generated response in the user's chat window.

[0291] Examples of prompt statements

[0292] Here are some example prompts to input to a generative AI model:

[0293] User: Why is it so slow?

[0294] Emotion engine: Anger recognition

[0295] Generative AI: We apologize for the inconvenience. We are currently working on improving the system.

[0296] In this way, a more appropriate response can be made according to the user's emotions.

[0297] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0298] Step 1:

[0299] Initial Server Setup

[0300] The server is installed within the company. The server administrator accesses the server's settings screen and sets filtering rules and security policies. Specifically, they set rules to block specific wording and apply security policies. The input is the administrator's settings information, and the output is the state of the server with the settings reflected.

[0301] Step 2:

[0302] User text input

[0303] A user uses a terminal to input text, specifically, typing "why is this so slow?" into a chat window. The input is the user's text, and the output is the transmission of text data from the terminal to the server.

[0304] Step 3:

[0305] Receiving and parsing text by the server

[0306] The server receives text input from the user. It analyzes the received text and checks whether it contains a specific phrase. Specifically, it detects the keyword "slow." The input is the user's text data, and the output is the analysis result.

[0307] Step 4:

[0308] Emotion recognition by emotion engine

[0309] The server sends the text to the emotion engine. The emotion engine analyzes the text and recognizes the user's emotion. Specifically, the emotion engine recognizes "anger." The input is the analyzed text data, and the output is the emotion recognition result.

[0310] Step 5:

[0311] Response generation using generative artificial intelligence

[0312] The server sends the emotion recognition results to the generative AI. The generative AI generates an appropriate response based on the emotion recognition results. Specifically, it generates a response such as "We apologize for the inconvenience. We are currently working on improving the system." The input is the emotion recognition results, and the output is the generated response.

[0313] Step 6:

[0314] Server sends response

[0315] The server sends the generated response to the user's terminal, and displays the generated response in a chat window. The input is the generated response data, and the output is the response displayed on the user's terminal.

[0316] (Application example 1)

[0317] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0318] Conventional generative AI systems were unable to respond in a way that took the user's emotions into account, and had insufficient functions for blocking specific phrases and for governance. This resulted in security risks and a poor user experience when used within a company. It was also difficult to provide appropriate responses when used on mobile devices such as smartphones.

[0319] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0320] In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific words, a means for recognizing the user's emotions using an emotion engine and adjusting the generative AI's response based on the emotion, and a means for functioning as an application installed on a smartphone. This makes it possible to reduce security risks within a company and provide appropriate responses according to the user's emotions.

[0321] The "Secure Proxy SaaS Service for Generative AI" is a Software as a Service (SaaS) that provides a proxy service for securely making inquiries to generative artificial intelligence (AI).

[0322] "Server" refers to a computer system that runs the secure proxy SaaS service for generative AI and processes and manages data within the company.

[0323] A "means of querying a generative AI via a server" is a method for sending data or requests to a generative AI via a server and receiving a response.

[0324] "Means of server management" refers to methods for managing the operation and security of servers and providing stable services.

[0325] "Measures to enable governance, such as blocking specific wording" refers to a method that provides the ability to detect and block specific keywords or phrases based on a company's security policy.

[0326] An "emotion engine" is a software component for analyzing emotions from a user's input text and generating an appropriate response based on the emotions.

[0327] "Means to function as an application installed on a smartphone" refers to software that operates on a smartphone and is provided in a form that the user can use.

[0328] The system for implementing this invention is composed mainly of a server that provides a secure proxy SaaS service for generative AI. The server is installed within a company and provides a proxy service for securely making inquiries to generative AI. The server utilizes operational and security knowledge cultivated in information and communications to enable governance such as blocking specific wording.

[0329] The server is equipped with an emotion engine that recognizes the user's emotions, receives text input from the user, analyzes the emotions from the text, and uses the TextBlob library to obtain the polarity of the emotions and adjusts the generative AI's response based on the user's emotions.

[0330] Users access the system through an application installed on their smartphones, which sends user input to the server and receives responses from it, which checks for specific phrases, generates prompts based on sentiment analysis, and uses the OpenAI API to generate appropriate responses.

[0331] For example, if a user enters "I'm very tired today," the emotion engine recognizes that the user is tired (a negative emotion). The server generates a prompt sentence like this:

[0332] "The user is tired. Generate an appropriate response to the following message: I'm very tired today."

[0333] Based on this prompt, a response is generated using the OpenAI API, and provided to the user, for example, "Thank you for your hard work. Please take it easy and get some rest today."

[0334] This system can reduce security risks within a company and provide appropriate responses according to the user's emotions.

[0335] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0336] Step 1:

[0337] A user starts an application installed on a smartphone and inputs text. The input text is a sentence in a natural language that includes the user's emotions and intentions. The input data is sent from the user's smartphone to a server.

[0338] Step 2:

[0339] The server parses the received text input and checks for specific terms. It checks whether the text contains any of the keywords or phrases in the list of specific terms. If the text contains any of the specific terms, the server generates a warning message and sends it back to the user. The input data is the text input, and the output data is either a warning message or text input for the next step.

[0340] Step 3:

[0341] If no specific phrase is included, the server performs sentiment analysis on the text input using the sentiment engine. It uses the TextBlob library to calculate the sentiment polarity (positive, negative, neutral) of the text. The input data is the text input, and the output data is the sentiment polarity.

[0342] Step 4:

[0343] The server generates a prompt sentence based on the result of the sentiment analysis. For example, if the sentiment polarity is negative, the server creates a prompt sentence in the form of "The user is tired. Please generate an appropriate response to the following message: [Text input]". The input data is the sentiment polarity and the text input, and the output data is the prompt sentence.

[0344] Step 5:

[0345] The server sends the generated prompt to the OpenAI API, which generates an appropriate response. The OpenAI API generates a natural language response based on the prompt and sends it back to the server. The input data is the prompt, and the output data is the generated response.

[0346] Step 6:

[0347] The server sends the generated response to the user's smartphone. The user receives and displays the response through an application on the smartphone. The input data is the generated response, and the output data is the response displayed on the user's smartphone.

[0348] Through the above processing steps, it is possible to provide an appropriate response according to the user's emotions.

[0349] Example 2

[0350] Next, a second embodiment of the second embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0351] Conventional systems using generative AI were unable to respond in a way that took the user's emotions into account, and there were also security concerns. In particular, when used within a company, there was a high risk of customer information and confidential information being leaked, and appropriate governance was required to prevent this. In addition, because it was not possible to generate appropriate responses based on the user's emotions, it was difficult to improve customer satisfaction.

[0352] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0353] In this invention, the server includes a means for providing a secure proxy service for the generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving text input from a user and sending the text to the generative AI, a means for the generative AI to recognize the user's emotions and generate a response based on the emotions, and a means for returning the generated response to the user. This makes it possible to generate an appropriate response corresponding to the user's emotions while ensuring security.

[0354] The "Secure Proxy Service for Generative AI" is a relay service that allows secure access to generative AI and has the ability to filter and encrypt data.

[0355] A "server" is a computer system that provides services to other computers over a network.

[0356] "Means of querying the generative AI via a server" refers to a function for sending data to the generative AI via a server and receiving the response.

[0357] "Server management" refers to the set of activities for operating, monitoring, and maintaining a server, including security measures and performance optimization.

[0358] "Governance such as blocking specific wording" is a management technique for preventing information leaks by detecting specific keywords or phrases and blocking or masking them.

[0359] "Means for receiving text input from a user and sending that text to the generative AI" refers to a function for receiving text data entered by a user and sending it to the generative AI.

[0360] "Means for the generative AI to recognize the user's emotions and generate a response based on those emotions" refers to the function that enables the generative AI to analyze emotions from the user's input text and generate an appropriate response based on the results of that analysis.

[0361] "Means of returning the generated response to the user" refers to a function for sending the response generated by the generative AI to the user via a server and displaying it to the user.

[0362] The present invention relates to a system for providing a secure proxy service for generative AI. A specific embodiment of this system will be described below.

[0363] System Configuration

[0364] server

[0365] The server is installed within the company and manages inquiries to the generative AI. The server utilizes operational and security knowledge cultivated in the information and communications industry to set up the system to block specific text (e.g., customer information or confidential information). The server uses the HTTPS protocol to encrypt data and ensure security.

[0366] Terminal

[0367] The terminal receives text input from the user and transmits the text to the server. The terminal is a device such as a smartphone or PC that interfaces with the user through a chat application.

[0368] Generative AI

[0369] The generative AI receives the user's text through the server and recognizes the user's emotions using an emotion engine. The generative AI analyzes the text using natural language processing (NLP) techniques and generates an appropriate response. The generated response is sent to the terminal via the server and displayed to the user.

[0370] Examples

[0371] Example 1: Customer Service

[0372] A user uses a terminal to type, "Tell me about your recent order." This text is sent from the terminal to a server, which filters out certain words. The filtered text is sent to a generative AI, which uses an emotion engine to recognize the user's emotions. For example, if a user types text containing the word "delayed," the generative AI recognizes the emotion of anger and generates a response that reads, "We apologize for the inconvenience. We will review the current status of your order." This response is sent to the terminal via the server and displayed to the user.

[0373] Examples of prompt statements

[0374] "If a user enters text that expresses anger, how should a generative AI respond?"

[0375] By feeding this prompt into a generative AI model, it can learn how to generate an appropriate response that corresponds to the user's emotions.

[0376] Hardware and Software Used

[0377] Server: A high-performance computer system (e.g. Linux server)

[0378] Devices: Smartphones, PCs

[0379] Generative AI: Generative AI models such as GPT-3

[0380] Protocol: HTTPS

[0381] Technology: Natural Language Processing (NLP), Emotion Recognition Engine

[0382] This system makes it possible to generate appropriate responses that correspond to the user's emotions while ensuring security.

[0383] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0384] Step 1:

[0385] A user uses the terminal to input text, for example, "Tell me about your recent orders." The input text is stored in the terminal's memory.

[0386] Step 2:

[0387] The terminal sends the text entered by the user to the server using the HTTPS protocol, and the data is encrypted with TLS. The input is the user's text, and the output is the encrypted text data.

[0388] Step 3:

[0389] The server filters the received text. If it contains certain words (e.g., customer information or confidential information), it blocks or masks that part. Regular expressions are used for filtering. The input is the encrypted text data, and the output is the filtered text data.

[0390] Step 4:

[0391] The server sends the filtered text to the generative AI, which includes an emotion engine to recognize the user's emotions. The input is the filtered text data, and the output is the emotion recognition result.

[0392] Step 5:

[0393] Generative AI recognizes the user's emotions from the received text. For example, if the text contains the word "delayed", it recognizes the emotion of anger. It uses natural language processing (NLP) techniques for emotion recognition. The input is filtered text data, and the output is the emotion recognition result.

[0394] Step 6:

[0395] Generative AI generates appropriate responses based on the recognized emotions. For example, if the user shows an angry emotion, it generates a calm and polite response. It uses a generative AI model such as GPT-3 to generate responses. The input is the emotion recognition result, and the output is the generated response.

[0396] Step 7:

[0397] The generative AI sends the generated response to the server. The server receives the response and sends it to the terminal. The input is the generated response, and the output is the encrypted response data.

[0398] Step 8:

[0399] The terminal displays the response it receives from the server to the user, for example, "We apologize for the inconvenience. We will check the current status of your order." The input is the encrypted response data, and the output is the text that is displayed to the user.

[0400] (Application example 2)

[0401] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0402] Conventional customer service systems using generative AI had problems such as difficulty in ensuring security and responding appropriately to customer emotions. In addition, there was a lack of concrete measures to streamline customer service in brick-and-mortar stores. This could lead to a decrease in customer satisfaction and the risk of information leaks.

[0403] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0404] In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for recognizing the user's emotions using an emotion engine and adjusting the generative AI's response based on the emotions, and a means for supporting customer service in a physical store using a smart device. This makes it possible to provide appropriate responses according to the customer's emotions while ensuring security, and to improve the efficiency of customer service in a physical store.

[0405] The "Secure Proxy SaaS Service for Generative AI" is a cloud-based service that ensures security when using generative AI and provides governance such as blocking specific wording.

[0406] A "server" is a computer system that processes queries to the generative AI and performs security management and data filtering.

[0407] A "means of querying a generative AI via a server" is a method for sending data or questions to a generative AI via a server and receiving a response.

[0408] "Means of server management" are methods for maintaining the operation and security of a server, and for implementing governance such as blocking specific text.

[0409] "Measures to enable governance, such as blocking specific wording" refers to a filtering function to remove confidential information and inappropriate wording from the data processed by the generative AI.

[0410] An "emotion engine" is software that analyzes text input from a user and recognizes the emotions contained in that text.

[0411] A "smart device" is a portable electronic device with Internet connectivity, such as a smartphone or tablet.

[0412] "Means to support customer service in physical stores" refer to technologies and methods for making customer service in physical stores more efficient and improving customer satisfaction.

[0413] This invention is a system for improving customer satisfaction by streamlining customer service in brick-and-mortar stores. This system is composed of a secure proxy SaaS service for generative AI, a server, an emotion engine, and a smart device.

[0414] System Configuration

[0415] 1. Secure proxy SaaS service for generative AI:

[0416] The server provides a secure proxy SaaS service for generative AI. This service ensures security when using generative AI and provides governance such as blocking specific wording.

[0417] 2. Server:

[0418] The server will be installed within the company and will handle queries to the generative AI. The server will be managed using operational and security knowledge cultivated in the information and communications industry.

[0419] 3. Emotion Engine:

[0420] The emotion engine analyzes text input from users, recognizes the emotions contained within that text, and adjusts the generative AI's response based on the recognized emotions.

[0421] 4. Smart Devices:

[0422] Smart devices (such as smartphones and tablets) are used to assist in customer service in brick-and-mortar stores, allowing store staff to respond quickly and appropriately to customer inquiries.

[0423] System Operation

[0424] The server receives customer inquiries from smart devices and uses an emotion engine to recognize the emotion contained in the inquiry. The recognized emotion and the inquiry content are sent to the generative AI, which generates an appropriate response. The generated response is then sent back to the smart device after filtering to block certain phrases.

[0425] Hardware and Software Used

[0426] Hardware: Servers, smartphones, tablets

[0427] Software: Secure proxy SaaS service for generative AI, emotion recognition API, generative AI API

[0428] Examples

[0429] For example, if a customer asks, "Can you tell me more about this product?", the emotion engine will recognize the customer's interest and the generative AI will provide a detailed product explanation. If a customer asks, "I would like to return a product," the emotion engine will recognize the customer's dissatisfaction and the generative AI will provide a polite response.

[0430] Examples of prompt statements

[0431] "Please tell me more about this product"

[0432] "I'd like to return it."

[0433] "Do you have this item in stock?"

[0434] In this way, a system for supporting customer service in brick-and-mortar stores can be realized. This system provides appropriate responses according to the customer's emotions while ensuring security, making it possible to improve the efficiency of customer service in brick-and-mortar stores.

[0435] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0436] Step 1:

[0437] A user enters a query using a smart device.

[0438] Input: User's query text (e.g. "Tell me more about this product")

[0439] Output: The query text is sent from the smart device to the server.

[0440] Step 2:

[0441] The server sends the received query text to an emotion engine to recognize emotions.

[0442] Input: Query text

[0443] Output: Perceived emotion (e.g. interest, anger, frustration)

[0444] Specific operation: The server calls the emotion recognition API and analyzes the query text to recognize emotions.

[0445] Step 3:

[0446] The server sends the recognized emotion and the query text to a generative AI to generate an appropriate response.

[0447] Input: query text, recognized sentiment

[0448] Output: The generated response text

[0449] How it works: The server calls the generative AI API, sends the query text and the recognized emotion as a prompt, and the generative AI generates a response.

[0450] Step 4:

[0451] The server filters the generated response text to block certain phrases.

[0452] Input: Generated response text

[0453] Output: The filtered response text

[0454] What happens: The server applies a filtering algorithm to filter out sensitive information and inappropriate language.

[0455] Step 5:

[0456] The server sends the filtered response text to the smart device.

[0457] Input: Filtered response text

[0458] Output: Response text displayed on smart device

[0459] Specific operation: The server sends the filtered response text to the smart device and displays it to the user.

[0460] Step 6:

[0461] The user checks the response on their smart device and makes additional enquiries if necessary.

[0462] Input: Filtered response text

[0463] Output: Additional query text from the user (e.g. "What other products do you recommend?")

[0464] Specific Actions: The user checks the response on the smart device and enters additional inquiries.

[0465] In this way, a system that supports customer service in brick-and-mortar stores is realized. The specific operations at each step ensure security and make it possible to provide appropriate responses according to the customer's emotions.

[0466] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires a voice indicating a user input for the result of the specific processing. The control unit 46A transmits the voice data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[0467] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by making a neural network perform deep learning. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating a voice, text data indicating a text, and image data indicating an image is input. The data generation model 58 performs inference on the input inference data according to the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0468] Another example of generative AI is Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) are mentioned.

[0469] In the above embodiment, an example was given in which the specific process was performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0470] [Second embodiment]

[0471] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0472] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0473] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a wide area network (WAN) and / or a local area network (LAN).

[0474] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0475] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs the voice according to instructions from the processor 46.

[0476] Camera 42 is a small digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (e.g., an imaging range defined by an angle of view equivalent to the width of the field of vision of an average healthy person).

[0477] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54. The transmission and reception of various types of information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is performed in a secure state.

[0478] Fig. 4 shows an example of main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0479] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32, and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0480] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0481] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0482] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0483] "Example 1"

[0484] The present invention is a system that provides a secure proxy SaaS service for generative AI. Specifically, a server is installed within the company, and queries are made to the generative AI via this server. Server management utilizes operational and security knowledge cultivated in the information and communications industry, enabling governance such as blocking specific wording through settings for each individual company.

[0485] "Example 2"

[0486] As a concrete example, if a company wants to use generative AI to respond to customers, they can use the system of the present invention to make inquiries to the generative AI via a server installed within the company. In this case, server management can utilize the operational and security knowledge cultivated in information and communications to set up settings to block specific text, such as customer information or confidential information. This makes it possible to use generative AI in business while ensuring security.

[0487] The process flow of each embodiment will be described below.

[0488] "Example 1"

[0489] Step 1: Set up a server within the company. This server acts as an intermediary for communication with the generative AI.

[0490] Step 2: A query is sent to the generative AI via the server. This query is based on the company's business needs, for example, generating an automatic response for customer support.

[0491] Step 3: Manage the server. This management utilizes operational and security knowledge cultivated in the information and communications industry, enabling governance such as blocking specific wording with settings for each individual company.

[0492] "Example 2"

[0493] Step 1: If a company wants to use generative AI to handle customer support, it first sets up a server within the company.

[0494] Step 2: Make an inquiry to the generative AI via the installed server. This inquiry may involve generating an automatic response to a customer inquiry.

[0495] Step 3: Manage the server. This management utilizes operational and security knowledge cultivated in the information and communications industry, and sets up settings to block specific text, such as customer information or confidential information. This makes it possible to use generative AI in business while ensuring security.

[0496] Example 1

[0497] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0498] In order to use generative AI models safely and efficiently within a company, it is necessary to minimize the risk of data leakage to the outside while implementing governance based on the company's security policy. However, conventional systems lack the ability to block or filter specific text, making it difficult to fully comply with a company's security policy.

[0499] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0500] In this invention, the server includes a means for providing a secure proxy service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving a prompt text from a user's terminal, a means for inspecting the received prompt text based on filtering rules, a means for relaying the prompt text that has passed the inspection to the generative AI model, a means for receiving a response from the generative AI model and inspecting it again based on filtering rules, and a means for transmitting the response that has passed the filtering to the user's terminal. This makes it possible to safely and efficiently use the generative AI model while realizing governance based on the company's security policy.

[0501] The "Secure Proxy Service for Generative AI" is a service that securely relays queries to generative AI models and enables governance based on a company's security policies.

[0502] A “server” is a computer system installed within a company that relays queries to the generative AI model and provides filtering and governance functions.

[0503] A "prompt" is text that contains instructions or questions that a user enters into a generative AI model.

[0504] "Filtering rules" are rules or conditions for blocking specific words or phrases, and are set based on a company's security policy.

[0505] A "generative AI model" is an artificial intelligence model that generates responses based on prompts from the user.

[0506] "Governance" refers to management and control, such as blocking specific words or phrases, based on a company's security policy.

[0507] A "user's device" is a device, such as a computer or smartphone, that a user uses to input prompts to a generative AI model and receive responses.

[0508] An "inquiry" refers to instructions or questions that a user gives to a generative AI model, and is sent to the server as a prompt.

[0509] A "response" is an answer or piece of information that a generative AI model generates based on a prompt.

[0510] This invention is a system that provides a secure proxy service for generative AI, and queries generative AI models through a server installed within a company. A specific embodiment of this system will be described below.

[0511] Server installation and initial configuration

[0512] The server is installed within the company and connected to the company's network. The server administrator accesses the server's settings screen and sets filtering rules based on the company's security policy. For example, a rule is set to block inquiries that include the phrase "confidential information." Once the settings are complete, the server is connected to the company's network and begins operation.

[0513] User query input

[0514] The user inputs a prompt for the generative AI model from their own device. For example, a marketing person inputs a prompt such as, "Please tell me your sales forecast for the next quarter." Once the input is complete, the user clicks the send button.

[0515] The server receives and examines the query

[0516] The server receives the prompt text sent from the user's terminal. It checks the received prompt text based on the filtering rules. For example, it checks whether the text contains the phrase "confidential information." If there are no problems, it proceeds to the next step.

[0517] Relaying inquiries

[0518] The server relays the prompts that pass inspection to the generative AI model, which generates a response based on the prompts received. For example, it generates a response like, "Sales forecast for the next quarter is XXX yen."

[0519] Receiving and filtering responses by the server

[0520] The server receives the response from the generative AI model. The server checks the response again based on the filtering rules. For example, it checks whether the response contains the phrase "confidential information." If there are no problems, it proceeds to the next step.

[0521] Providing a response

[0522] The server sends the filtered responses to the user's device. The user can check the response from the generative AI model on the device. For example, the user can check the response, "Sales forecast for the next quarter is XXX yen."

[0523] Examples

[0524] As a concrete example, Company A has a setting to block queries that contain the phrase "confidential information." When a user enters the prompt phrase "Please generate a report that contains confidential information," the server detects this phrase and returns an error message without relaying the query.

[0525] Company B also uses a generative AI model to generate marketing reports. When a user enters a prompt such as "Please generate a report on the latest market trends," the server inspects the query and relays it to the generative AI model if it determines there is no problem. The response from the generative AI model is provided to the user via the server.

[0526] The system enables companies to use generative AI models safely and efficiently.

[0527] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0528] Step 1:

[0529] The server is installed within the company and connected to the network. The server administrator accesses the server's configuration screen and sets filtering rules based on the company's security policy. For example, a rule is set to block inquiries that include the phrase "confidential information." The input is the security policy and filtering rules, and the output is the configured server.

[0530] Step 2:

[0531] The user inputs a prompt sentence to the generative AI model from his / her own device. For example, the user inputs a prompt sentence such as "Please tell me your sales forecast for the next quarter." The input is the prompt sentence, and the output is the sent prompt sentence. The user clicks the send button.

[0532] Step 3:

[0533] The server receives the prompt text sent from the user's terminal. It checks the received prompt text based on filtering rules. For example, it checks whether it contains the phrase "confidential information." The input is the sent prompt text, and the output is the inspection result. If there are no problems, it proceeds to the next step.

[0534] Step 4:

[0535] The server relays the prompts that pass inspection to the generative AI model. The generative AI model generates a response based on the prompts it receives. For example, it generates a response like, "Sales forecast for the next quarter is XXX yen." The input is the prompt that passes inspection, and the output is the generated response.

[0536] Step 5:

[0537] The server receives the response from the generative AI model. The received response is inspected again based on the filtering rules. For example, it checks whether it contains the phrase "confidential information." The input is the response from the generative AI model, and the output is the inspection result. If there are no problems, it proceeds to the next step.

[0538] Step 6:

[0539] The server sends the filtered response to the user's device. The user can check the response from the generative AI model on the device. For example, the user can check the response "Sales forecast for the next quarter is XXX yen." The input is the filtered response, and the output is the response sent to the user's device.

[0540] (Application example 1)

[0541] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0542] When using generative AI within a company, there is a risk that confidential information may be leaked to the outside. In addition, it is difficult to implement appropriate governance when specific wording or data is included. Furthermore, security must be ensured when making inquiries to generative AI using smartphones.

[0543] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means. In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing a server, a means for making an inquiry to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for making an inquiry to the generative AI through an application installed on a smartphone, a means for securely communicating via a secure proxy server, a means for sending a filtered prompt to the generative AI, and a means for displaying the result on a smartphone application. This makes it possible to reduce the risk of confidential information leakage and perform appropriate governance when using generative AI within a company. In addition, high security can be ensured even when making inquiries to the generative AI using a smartphone.

[0544] The "Secure Proxy SaaS Service for Generative AI" is a cloud-based service that provides a proxy server for securely making inquiries to generative AI.

[0545] A "server" is a computer system that relays queries to the generative AI and manages security and governance.

[0546] A "means for querying a generative AI via a server" is a method for sending a prompt to a generative AI via a server and receiving a result.

[0547] "Means of server management" refers to the methods for maintaining and properly managing the operation and security of the server.

[0548] "Measures that enable governance, such as blocking specific wording" refers to methods for filtering out specific keywords and phrases and controlling the content of queries made to generative AI.

[0549] An "application installed on a smartphone" is software that runs on a smartphone and is used to make queries to the generative AI.

[0550] "Means for securely communicating via a secure proxy server" refers to a method for encrypting data and communicating securely through a secure proxy server.

[0551] "Means for transmitting a filtered prompt to the generative AI" refers to a method for transmitting the prompt after it has been filtered to the generative AI.

[0552] "Means for displaying results in a smartphone application" refers to a method for displaying the results returned from the generative AI in an application on a smartphone.

[0553] The following system configuration will be described as an embodiment of the present invention.

[0554] System Configuration

[0555] Hardware

[0556] Server: A computer system installed within a company that provides a secure proxy SaaS service for generative AI.

[0557] Smartphone: A device that allows users to make inquiries to generative AI.

[0558] software

[0559] Secure Proxy Server: Runs on the server, relays queries to the generative AI, and manages security and governance.

[0560] Smartphone application: Software that is installed on a smartphone and allows for queries to be made to the generative AI.

[0561] Data processing and calculation

[0562] server

[0563] The server provides a secure proxy SaaS service for generative AI. When a user makes a query to generative AI through a smartphone application, the prompt is first sent to the server. The server applies governance such as blocking certain words, and sends the filtered prompt to the generative AI.

[0564] Smartphone

[0565] A user uses a smartphone application to make a query to the generative AI. After entering a prompt and pressing the send button, the prompt is sent to the generative AI via a secure proxy server. The results returned from the generative AI are then displayed on the smartphone application via the secure proxy server again.

[0566] Examples

[0567] For example, when a company queries a generative AI for documents containing confidential information, the system can be used to reduce the risk of confidential information leaking to the outside. The user enters the following prompt into a smartphone application:

[0568] Prompt: "Generate the following sentence: Details about a company's new product."

[0569] This prompt is sent to a server where it is filtered to see if it contains specific words, and the filtered prompt is sent to a generative AI, which then displays the generated results on a smartphone application.

[0570] In this way, it is possible to reduce the risk of confidential information leaks and implement appropriate governance when using generative AI within a company. It is also possible to ensure high security when making inquiries to generative AI using smartphones.

[0571] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0572] Step 1:

[0573] The user starts the smartphone application and inputs the question (prompt) they want to ask the generative AI. The input prompt is temporarily saved in the application.

[0574] Step 2:

[0575] When the user presses the send button, the smartphone application sends the prompt text to the secure proxy server. The input is the prompt text, and the output is a send request to the secure proxy server.

[0576] Step 3:

[0577] The server analyzes the received prompt text and filters it to see if it contains specific words. The input is the prompt text, and the output is the filtered prompt text. Specifically, the server checks it against a specific keyword list and removes any inappropriate words that it finds.

[0578] Step 4:

[0579] The server sends the filtered prompt to the generative AI. The input is the filtered prompt, and the output is a request to send to the generative AI. Specifically, the server sends an HTTP request to the API endpoint of the generative AI.

[0580] Step 5:

[0581] The generative AI generates a response based on the received prompt and sends the result back to the server. The input is the filtered prompt and the output is the generated response. In concrete terms, the generative AI uses its internal model to generate a response to the prompt.

[0582] Step 6:

[0583] The server receives the response returned from the generative AI and sends it to the smartphone application. The input is the generated response, and the output is a request to send to the smartphone application. In concrete terms, the server sends an HTTP request with the response data to the API endpoint of the smartphone application.

[0584] Step 7:

[0585] The smartphone application displays the response received from the server. The input is the generated response and the output is the result displayed to the user. In concrete terms, the application passes the received data to a UI component for display on the screen.

[0586] Example 2

[0587] Next, a description will be given of Example 2 of the embodiment 2. In the following description, the data processing device 12 is referred to as a "server" and the smart glasses 214 are referred to as a "terminal".

[0588] Customer support systems that use generative AI are required to respond to customers efficiently while ensuring security. However, with conventional systems, there is a risk that customer information or confidential information may be sent to the generative AI model, which may lead to security issues. In addition, responses from the generative AI model may be difficult for users to understand, which can lead to a decline in the quality of customer support. To solve these issues, a system that can filter the content of inquiries and convert the content of responses is needed.

[0589] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0590] In this invention, the server includes a means for receiving an inquiry from a terminal, a means for analyzing the received inquiry content and filtering out specific wording, a means for generating a prompt sentence based on the filtered inquiry content, a means for sending the generated prompt sentence to a generative AI model, a means for receiving a response from the generative AI model, a means for analyzing the received response and converting it into a format that is easy for a user to understand, a means for sending the converted response to the terminal, a means for installing a server, a means for making an inquiry to the generative AI model via the server, a means for managing the server, and a means for enabling governance such as blocking specific wording. This enables efficient customer service using generative AI while ensuring security.

[0591] A "terminal" is a device through which a user inputs inquiry content and transmits it to a server.

[0592] A "server" is a device that receives queries from a terminal, queries the generative AI model, and sends responses to the terminal.

[0593] "Inquiry content" is text data of a question or request that a user inputs via a terminal.

[0594] "Filtering" is the process of removing or masking specific wording (such as customer information or confidential information) from the content of an inquiry.

[0595] A "prompt sentence" is the text of a query that is generated to be sent to a generative AI model.

[0596] A "generative AI model" is an artificial intelligence model that generates a response based on an input prompt sentence.

[0597] A "response" is the text data of the answer returned from the generative AI model.

[0598] "Analysis" is the process of understanding the inquiry and response, and extracting or converting the necessary information.

[0599] "Transformation" is the process of changing the response from a generative AI model into a form that is easy for a user to understand.

[0600] "Governance" refers to the means of managing the operation and security of a system, and exercising control such as blocking specific text.

[0601] This invention relates to a customer service system that uses generative AI. Specifically, it is a system that uses a terminal, a server, and a generative AI model to filter inquiries and convert responses.

[0602] Hardware and Software Used

[0603] The terminal is a device where a user inputs the contents of an inquiry and transmits the input to the server. The terminal may be a PC, a smartphone, a tablet, or the like.

[0604] The server is a device that receives queries from the terminal, queries the generative AI model, and sends the response to the terminal. The server has installed software for parsing the query content, filtering, generating prompts, and parsing and converting the response. Specific software includes a natural language processing (NLP) library and an API client.

[0605] A generative AI model is an artificial intelligence model that generates responses based on input prompt sentences. Specifically, OpenAI's GPT-4 and other models are used.

[0606] Data processing and calculation

[0607] The terminal transmits the inquiry content input by the user to the server. For example, the user inputs "Please tell me the price of the new product."

[0608] The server analyzes the received inquiry content and filters out specific phrases (such as customer information or confidential information). Based on the filtered inquiry content, it generates a prompt sentence suitable for the generative AI model. For example, it generates a prompt sentence such as "Please tell me the price of the new product."

[0609] The server sends the generated prompt sentence to the generative AI model, which generates a response based on the prompt sentence and sends it back to the server. For example, it generates a response such as "The price of the new product is $100."

[0610] The server parses the response and converts it into a format that is easy for the user to understand. It then sends the converted response to the terminal. The terminal displays the response to the user. For example, the message "The price of the new product is $100" is displayed on the user's screen.

[0611] Examples

[0612] For example, if a company receives an inquiry from a customer saying, "Please tell me the price of a new product," the following process will take place:

[0613] 1. On the terminal, the user enters "Please tell me the price of the new product" and clicks the send button.

[0614] 2. The server analyzes the received data and extracts the text "Please tell me the price of the new product." It then performs a filtering process to check whether the data contains a specific phrase.

[0615] 3. The server generates a prompt sentence based on the filtered text: "Please tell me the price of the new product."

[0616] 4. The server creates and sends an API request to send the generated prompt to GPT-4.

[0617] 5. The server receives the response from GPT-4 and gets the text “The price of the new product is $100.”

[0618] 6. The server parses the received text and replaces technical terms with more common language where necessary.

[0619] 7. The server sends the converted response to the terminal, which displays the received response to the user, and the message "The new product costs $100" appears on the user's screen.

[0620] In this way, businesses can leverage generative AI to deliver efficient customer service while ensuring security.

[0621] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0622] Step 1:

[0623] The terminal transmits the inquiry content entered by the user to the server. As input, the user inputs "Please tell me the price of the new product" and clicks the send button. As output, the inquiry content is transmitted to the server.

[0624] Step 2:

[0625] The server analyzes the inquiry received from the terminal. As input, it receives the text data "Please tell me the price of the new product" sent from the terminal. To process the data, it analyzes the text using natural language processing (NLP) technology and filters out specific phrases (such as customer information or confidential information). As output, it obtains the filtered inquiry content.

[0626] Step 3:

[0627] The server generates a prompt sentence based on the filtered inquiry content. As input, it receives the filtered text "Please tell me the price of the new product." As data processing, it converts it into a format suitable for the generative AI model. As output, it obtains the generated prompt sentence.

[0628] Step 4:

[0629] The server sends the generated prompt sentence to the generative AI model. As input, it receives the generated prompt sentence "Please tell me the price of the new product." As a specific operation, it creates an API request and sends it to the generative AI model (e.g., GPT-4). As output, the prompt sentence is sent to the generative AI model.

[0630] Step 5:

[0631] The server receives the response from the generative AI model. As input, it receives the response data returned from the generative AI model. As a specific operation, it parses the API response and obtains the text "The price of the new product is $100." As output, it obtains the response from the generative AI model.

[0632] Step 6:

[0633] The server analyzes the received response and converts it into a format that is easy for the user to understand. As input, it receives the response data "The price of the new product is $100." As data processing, it performs conversion processing such as replacing technical terms with common terms. As output, it obtains the converted response.

[0634] Step 7:

[0635] The server sends the converted response to the terminal. As input, it receives the converted response data "The price of the new product is $100." As a specific operation, it sends the response data to the terminal. As output, the response is sent to the terminal.

[0636] Step 8:

[0637] The terminal displays the received response to the user. As input, it receives the response data "The price of the new product is $100" sent from the server. As a specific operation, it displays the response message on the user's screen. As output, the user can confirm the response.

[0638] (Application example 2)

[0639] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the smart glasses 214 will be referred to as a "terminal".

[0640] Conventional customer service systems using generative AI have difficulty in ensuring security and cannot block certain text (e.g., customer information or confidential information), making it difficult to use safely within a company. In addition, they lacked the functionality to make personalized product recommendations to users, making it difficult to improve customer satisfaction.

[0641] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0642] In this invention, the server includes a means for providing a secure proxy service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving questions from users and sending them to the generative AI, a means for returning answers from the generative AI to the user, and a means for making personalized product recommendations based on the user's purchase history and browsing history. This enables safe customer service and personalized product recommendations using generative AI while ensuring security.

[0643] The "Secure Proxy Service for Generative AI" is a relay service that enables safe inquiries to be made to generative AI, and exchanges data while ensuring security.

[0644] A "server" is a computer system that stores, processes, and manages data on a network.

[0645] A "means for querying a generative AI via a server" is a method for sending questions or instructions to a generative AI via a server.

[0646] "Server management" refers to the activity of operating, maintaining, and managing the security of a server.

[0647] "Governance such as blocking specific wording" is a management technique that filters out specific keywords and phrases (such as customer information or confidential information) to prevent the leakage of inappropriate information within the system.

[0648] "Means for receiving questions from a user and sending them to a generative AI" refers to a method for receiving questions entered by a user and sending them to a generative AI.

[0649] "Means of returning an answer from a generative AI to a user" refers to a method for returning an answer obtained from a generative AI to a user.

[0650] "Means for personalized product recommendations based on a user's purchase history and browsing history" refers to a method for analyzing a user's past purchase history and browsing history and recommending individually optimized products based on that.

[0651] A system for implementing this invention includes a server that provides a secure proxy service for the generative AI, a terminal that receives questions from users and sends them to the generative AI, a terminal that returns answers from the generative AI to the user, and a function that makes personalized product recommendations based on the user's purchasing history and browsing history.

[0652] Hardware and Software Configuration

[0653] Server: Installed within a company, it provides a secure proxy service for generative AI. The server is built using a web framework such as Flask.

[0654] Device: A device used by a user, such as a smartphone or tablet. These devices run applications that receive questions from users and send them to the generative AI.

[0655] Generative AI: Use generative AI models such as OpenAI's GPT-3. Generative AI generates appropriate answers to questions posed by users.

[0656] Data processing and calculation

[0657] 1. Receiving user questions:

[0658] The user inputs a question through the terminal, for example, "How do you use this product?"

[0659] 2. Filter questions:

[0660] The server filters certain phrases (customer information and confidential information) from the question, ensuring security.

[0661] 3. Querying the generative AI:

[0662] The filtered questions are sent to a generative AI, which generates answers to the questions.

[0663] 4. Response to the answer:

[0664] The server receives the answer from the generative AI and responds to the user through the device, for example, "This product is used in the following steps..."

[0665] 5. Personalized product recommendations:

[0666] The server analyzes the user's purchase and browsing history and recommends individually optimized products based on that information, for example by displaying "These are the products we recommend for you."

[0667] Examples

[0668] When a user asks, "How do you use this product?", the server filters the question and sends it to the generative AI. The generative AI responds with, "This product is used in the following steps..." and the answer is returned to the user. Related products are also recommended based on the user's past purchase and browsing history.

[0669] Examples of prompt statements

[0670] "How do you use this product?"

[0671] In this way, secure customer interactions and personalized product recommendations using generative AI can be achieved while ensuring security.

[0672] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0673] Step 1:

[0674] A user inputs a question through a terminal, for example, "How do you use this product?" The input question is sent to a server by an application on the terminal.

[0675] Step 2:

[0676] The server filters the questions it receives. Specifically, it detects specific phrases (customer information or confidential information) from the questions and blocks or masks them. The filtered questions are then sent to the generative AI.

[0677] Step 3:

[0678] The server sends the filtered questions to the generative AI, which generates appropriate answers to the questions it receives. For example, it uses OpenAI's GPT-3 to generate an answer like, "This product is used in the following steps..."

[0679] Step 4:

[0680] The server receives the answer from the generative AI. The server returns the answer to the device. The device displays the answer from the generative AI to the user.

[0681] Step 5:

[0682] The server analyzes the user's purchase history and browsing history. Specifically, it retrieves the user's past purchase history and browsing history from the database and makes personalized product recommendations based on that. For example, it displays "These are the products we recommend for you."

[0683] Step 6:

[0684] The server sends the personalized product recommendations to the terminal. The terminal displays the recommended products to the user. The user can review the recommended products and consider purchasing them.

[0685] In this way, secure customer interactions and personalized product recommendations using generative AI can be achieved while ensuring security.

[0686] In addition, an emotion engine that estimates the emotion of the user may be further combined. That is, the identification processing unit 290 may estimate the emotion of the user using the emotion identification model 59, and perform identification processing using the emotion of the user.

[0687] "Example 1"

[0688] As one embodiment of the present invention, there is a system that provides a secure proxy SaaS service for generative AI, which is combined with an emotion engine. This system queries the generative AI through a server installed in the company and manages the server. The server management utilizes operational and security knowledge cultivated in information and communications, enabling governance such as blocking specific wording. Furthermore, the generative AI includes an emotion engine that recognizes the user's emotions. Specifically, it receives text input from the user and recognizes the user's emotions from the text. Then, the generative AI's response is adjusted based on the recognized emotion. For example, if a user inputs text indicating an emotion of anger, the emotion engine recognizes the anger, and the generative AI generates an appropriate response based on the recognition result. This enables a more appropriate response that corresponds to the user's emotions.

[0689] "Example 2"

[0690] As one embodiment of the present invention, there is a system that provides a secure proxy SaaS service for generative AI, which is combined with an emotion engine. This system queries the generative AI through a server installed in the company and manages the server. The server management utilizes operational and security knowledge cultivated in information and communications, enabling governance such as blocking specific wording. Furthermore, the generative AI includes an emotion engine that recognizes the user's emotions. Specifically, it receives text input from the user and recognizes the user's emotions from the text. Then, the generative AI's response is adjusted based on the recognized emotion. For example, if a user inputs text indicating an emotion of anger, the emotion engine recognizes the anger, and the generative AI generates an appropriate response based on the recognition result. This enables a more appropriate response that corresponds to the user's emotions.

[0691] The process flow of each embodiment will be described below.

[0692] "Example 1"

[0693] Step 1: Receive text input from a user, which may include the user's sentiment.

[0694] Step 2: Input the received text into the emotion engine, which recognizes the user's emotion from the text.

[0695] Step 3: The output from the emotion engine (the recognized emotion) is fed into the generative AI, which generates a response based on this emotion.

[0696] Step 4: The generative AI's response is returned to the user. This response corresponds to the user's emotion.

[0697] Example 1

[0698] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0699] Conventional services using generative AI have security and governance issues, making it difficult to block specific text or operate based on a company's security policy. In addition, they are unable to generate responses that take into account the user's emotions, which can lead to a poor user experience.

[0700] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0701] In this invention, the server includes a means for providing a secure proxy service for the generative AI, a means for installing the server, a means for making an inquiry to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific words, a means for including an emotion engine for recognizing the user's emotions, and a means for adjusting the response of the generative AI based on the emotions recognized by the emotion engine. This makes it possible to generate an appropriate response corresponding to the user's emotions while strengthening security and governance.

[0702] "Generative artificial intelligence" is an artificial intelligence technology that generates natural language responses and content based on user input.

[0703] A "secure proxy service" is a service that relays user requests and provides proxy functions to enhance security and governance.

[0704] A "server" is a computer system that processes and manages data on a network.

[0705] "Governance" refers to the management and control of system operations, such as blocking specific wording and applying security policies.

[0706] An "emotion engine" is software or hardware for analyzing text input from a user and recognizing its emotions.

[0707] "Adjusting the response" means changing the content of the response generated by the generative artificial intelligence based on the emotion recognized by the emotion engine.

[0708] The present invention is a system that provides a secure proxy service for querying a generative artificial intelligence through a server installed within a company. The system includes an emotion engine that recognizes the user's emotions and adjusts responses based on the emotions.

[0709] Hardware and Software Used

[0710] 1. Server

[0711] It is installed within a company.

[0712] It will be managed utilizing operational and security knowledge cultivated in the information and communications industry.

[0713] It has a governance function that blocks specific wording.

[0714] 2. Terminal

[0715] A device (e.g. PC, smartphone) that allows a user to input text.

[0716] 3. Generative Artificial Intelligence

[0717] An artificial intelligence technology that generates natural language responses based on user input.

[0718] 4. Emotion Engine

[0719] Software that analyzes text input from users and recognizes their emotions.

[0720] Data processing and calculation

[0721] 1. User text input

[0722] A user uses the terminal to input text.

[0723] For example: "Why is it so slow?"

[0724] 2. Receiving and parsing the text by the server

[0725] The server receives text input from the user and checks whether it contains specific words.

[0726] Example: Detect the keyword "slow".

[0727] 3. Emotion Recognition by Emotion Engine

[0728] The server sends the text to an emotion engine, which analyzes the text to recognize the user's emotion.

[0729] Example: The emotion engine recognizes "anger."

[0730] 4. Response generation using generative artificial intelligence

[0731] The server sends the emotion recognition results to the generative artificial intelligence, which then generates an appropriate response.

[0732] For example: "We apologize for the inconvenience. We are currently working on improving our system."

[0733] 5. Server Sends Response

[0734] The server sends the generated response to the user's terminal.

[0735] Examples

[0736] For example, if a user enters the text "Why is it so slow?", the following happens:

[0737] 1. User text input

[0738] A user types in the chat window, "Why is this so slow?"

[0739] 2. Receiving and parsing the text by the server

[0740] The server receives the text "Why is it so slow?" and detects the keyword "slow".

[0741] 3. Emotion Recognition by Emotion Engine

[0742] The server sends the text to an emotion engine, which recognizes "anger."

[0743] 4. Response generation using generative artificial intelligence

[0744] The server sends the emotion recognition results to the generative artificial intelligence, which generates a response saying, "We apologize for the inconvenience. We are currently working on improving the system."

[0745] 5. Server Sends Response

[0746] The server displays the generated response in the user's chat window.

[0747] Examples of prompt statements

[0748] Here are some example prompts to input to a generative AI model:

[0749] User: Why is it so slow?

[0750] Emotion engine: Anger recognition

[0751] Generative AI: We apologize for the inconvenience. We are currently working on improving the system.

[0752] In this way, a more appropriate response can be made according to the user's emotions.

[0753] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0754] Step 1:

[0755] Initial Server Setup

[0756] The server is installed within the company. The server administrator accesses the server's settings screen and sets filtering rules and security policies. Specifically, they set rules to block specific wording and apply security policies. The input is the administrator's settings information, and the output is the state of the server with the settings reflected.

[0757] Step 2:

[0758] User text input

[0759] A user uses a terminal to input text, specifically, typing "why is this so slow?" into a chat window. The input is the user's text, and the output is the transmission of text data from the terminal to the server.

[0760] Step 3:

[0761] Receiving and parsing text by the server

[0762] The server receives text input from the user. It analyzes the received text and checks whether it contains a specific phrase. Specifically, it detects the keyword "slow." The input is the user's text data, and the output is the analysis result.

[0763] Step 4:

[0764] Emotion recognition by emotion engine

[0765] The server sends the text to the emotion engine. The emotion engine analyzes the text and recognizes the user's emotion. Specifically, the emotion engine recognizes "anger." The input is the analyzed text data, and the output is the emotion recognition result.

[0766] Step 5:

[0767] Response generation using generative artificial intelligence

[0768] The server sends the emotion recognition results to the generative AI. The generative AI generates an appropriate response based on the emotion recognition results. Specifically, it generates a response such as "We apologize for the inconvenience. We are currently working on improving the system." The input is the emotion recognition results, and the output is the generated response.

[0769] Step 6:

[0770] Server sends response

[0771] The server sends the generated response to the user's terminal, and displays the generated response in a chat window. The input is the generated response data, and the output is the response displayed on the user's terminal.

[0772] (Application example 1)

[0773] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0774] Conventional generative AI systems were unable to respond in a way that took the user's emotions into account, and had insufficient functions for blocking specific phrases and for governance. This resulted in security risks and a poor user experience when used within a company. It was also difficult to provide appropriate responses when used on mobile devices such as smartphones.

[0775] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0776] In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific words, a means for recognizing the user's emotions using an emotion engine and adjusting the generative AI's response based on the emotion, and a means for functioning as an application installed on a smartphone. This makes it possible to reduce security risks within a company and provide appropriate responses according to the user's emotions.

[0777] The "Secure Proxy SaaS Service for Generative AI" is a Software as a Service (SaaS) that provides a proxy service for securely making inquiries to generative artificial intelligence (AI).

[0778] "Server" refers to a computer system that runs the secure proxy SaaS service for generative AI and processes and manages data within the company.

[0779] A "means of querying a generative AI via a server" is a method for sending data or requests to a generative AI via a server and receiving a response.

[0780] "Means of server management" refers to methods for managing the operation and security of servers and providing stable services.

[0781] "Measures to enable governance, such as blocking specific wording" refers to a method that provides the ability to detect and block specific keywords or phrases based on a company's security policy.

[0782] An "emotion engine" is a software component for analyzing emotions from a user's input text and generating an appropriate response based on the emotions.

[0783] "Means to function as an application installed on a smartphone" refers to software that operates on a smartphone and is provided in a form that the user can use.

[0784] The system for implementing this invention is composed mainly of a server that provides a secure proxy SaaS service for generative AI. The server is installed within a company and provides a proxy service for securely making inquiries to generative AI. The server utilizes operational and security knowledge cultivated in information and communications to enable governance such as blocking specific wording.

[0785] The server is equipped with an emotion engine that recognizes the user's emotions, receives text input from the user, analyzes the emotions from the text, and uses the TextBlob library to obtain the polarity of the emotions and adjusts the generative AI's response based on the user's emotions.

[0786] Users access the system through an application installed on their smartphones, which sends user input to the server and receives responses from it, which checks for specific phrases, generates prompts based on sentiment analysis, and uses the OpenAI API to generate appropriate responses.

[0787] For example, if a user enters "I'm very tired today," the emotion engine recognizes that the user is tired (a negative emotion). The server generates a prompt sentence like this:

[0788] "The user is tired. Generate an appropriate response to the following message: I'm very tired today."

[0789] Based on this prompt, a response is generated using the OpenAI API, and provided to the user, for example, "Thank you for your hard work. Please take it easy and get some rest today."

[0790] This system can reduce security risks within a company and provide appropriate responses according to the user's emotions.

[0791] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0792] Step 1:

[0793] A user starts an application installed on a smartphone and inputs text. The input text is a sentence in a natural language that includes the user's emotions and intentions. The input data is sent from the user's smartphone to a server.

[0794] Step 2:

[0795] The server parses the received text input and checks for specific terms. It checks whether the text contains any of the keywords or phrases in the list of specific terms. If the text contains any of the specific terms, the server generates a warning message and sends it back to the user. The input data is the text input, and the output data is either a warning message or text input for the next step.

[0796] Step 3:

[0797] If no specific phrase is included, the server performs sentiment analysis on the text input using the sentiment engine. It uses the TextBlob library to calculate the sentiment polarity (positive, negative, neutral) of the text. The input data is the text input, and the output data is the sentiment polarity.

[0798] Step 4:

[0799] The server generates a prompt sentence based on the result of the sentiment analysis. For example, if the sentiment polarity is negative, the server creates a prompt sentence in the form of "The user is tired. Please generate an appropriate response to the following message: [Text input]". The input data is the sentiment polarity and the text input, and the output data is the prompt sentence.

[0800] Step 5:

[0801] The server sends the generated prompt to the OpenAI API, which generates an appropriate response. The OpenAI API generates a natural language response based on the prompt and sends it back to the server. The input data is the prompt, and the output data is the generated response.

[0802] Step 6:

[0803] The server sends the generated response to the user's smartphone. The user receives and displays the response through an application on the smartphone. The input data is the generated response, and the output data is the response displayed on the user's smartphone.

[0804] Through the above processing steps, it is possible to provide an appropriate response according to the user's emotions.

[0805] Example 2

[0806] Next, a description will be given of Example 2 of the embodiment 2. In the following description, the data processing device 12 is referred to as a "server" and the smart glasses 214 are referred to as a "terminal".

[0807] Conventional systems using generative AI were unable to respond in a way that took the user's emotions into account, and there were also security concerns. In particular, when used within a company, there was a high risk of customer information and confidential information being leaked, and appropriate governance was required to prevent this. In addition, because it was not possible to generate appropriate responses based on the user's emotions, it was difficult to improve customer satisfaction.

[0808] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0809] In this invention, the server includes a means for providing a secure proxy service for the generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving text input from a user and sending the text to the generative AI, a means for the generative AI to recognize the user's emotions and generate a response based on the emotions, and a means for returning the generated response to the user. This makes it possible to generate an appropriate response corresponding to the user's emotions while ensuring security.

[0810] The "Secure Proxy Service for Generative AI" is a relay service that allows secure access to generative AI and has the ability to filter and encrypt data.

[0811] A "server" is a computer system that provides services to other computers over a network.

[0812] "Means of querying the generative AI via a server" refers to a function for sending data to the generative AI via a server and receiving the response.

[0813] "Server management" refers to the set of activities for operating, monitoring, and maintaining a server, including security measures and performance optimization.

[0814] "Governance such as blocking specific wording" is a management technique for preventing information leaks by detecting specific keywords or phrases and blocking or masking them.

[0815] "Means for receiving text input from a user and sending that text to the generative AI" refers to a function for receiving text data entered by a user and sending it to the generative AI.

[0816] "Means for the generative AI to recognize the user's emotions and generate a response based on those emotions" refers to the function that enables the generative AI to analyze emotions from the user's input text and generate an appropriate response based on the results of that analysis.

[0817] "Means of returning the generated response to the user" refers to a function for sending the response generated by the generative AI to the user via a server and displaying it to the user.

[0818] The present invention relates to a system for providing a secure proxy service for generative AI. A specific embodiment of this system will be described below.

[0819] System Configuration

[0820] server

[0821] The server is installed within the company and manages inquiries to the generative AI. The server utilizes operational and security knowledge cultivated in the information and communications industry to set up the system to block specific text (e.g., customer information or confidential information). The server uses the HTTPS protocol to encrypt data and ensure security.

[0822] Terminal

[0823] The terminal receives text input from the user and transmits the text to the server. The terminal is a device such as a smartphone or PC that interfaces with the user through a chat application.

[0824] Generative AI

[0825] The generative AI receives the user's text through the server and recognizes the user's emotions using an emotion engine. The generative AI analyzes the text using natural language processing (NLP) techniques and generates an appropriate response. The generated response is sent to the terminal via the server and displayed to the user.

[0826] Examples

[0827] Example 1: Customer Service

[0828] A user uses a terminal to type, "Tell me about your recent order." This text is sent from the terminal to a server, which filters out certain words. The filtered text is sent to a generative AI, which uses an emotion engine to recognize the user's emotions. For example, if a user types text containing the word "delayed," the generative AI recognizes the emotion of anger and generates a response that reads, "We apologize for the inconvenience. We will review the current status of your order." This response is sent to the terminal via the server and displayed to the user.

[0829] Examples of prompt statements

[0830] "If a user enters text that expresses anger, how should a generative AI respond?"

[0831] By feeding this prompt into a generative AI model, it can learn how to generate an appropriate response that corresponds to the user's emotions.

[0832] Hardware and Software Used

[0833] Server: A powerful computer system (e.g. Linux server)

[0834] Devices: Smartphones, PCs

[0835] Generative AI: Generative AI models such as GPT-3

[0836] Protocol: HTTPS

[0837] Technology: Natural Language Processing (NLP), Emotion Recognition Engine

[0838] This system makes it possible to generate appropriate responses that correspond to the user's emotions while ensuring security.

[0839] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0840] Step 1:

[0841] A user uses the terminal to input text, for example, "Tell me about your recent orders." The input text is stored in the terminal's memory.

[0842] Step 2:

[0843] The terminal sends the text entered by the user to the server using the HTTPS protocol, and the data is encrypted with TLS. The input is the user's text, and the output is the encrypted text data.

[0844] Step 3:

[0845] The server filters the received text. If it contains certain words (e.g., customer information or confidential information), it blocks or masks that part. Regular expressions are used for filtering. The input is the encrypted text data, and the output is the filtered text data.

[0846] Step 4:

[0847] The server sends the filtered text to the generative AI, which includes an emotion engine to recognize the user's emotions. The input is the filtered text data, and the output is the emotion recognition result.

[0848] Step 5:

[0849] Generative AI recognizes the user's emotions from the received text. For example, if the text contains the word "delayed", it recognizes the emotion of anger. It uses natural language processing (NLP) techniques for emotion recognition. The input is filtered text data, and the output is the emotion recognition result.

[0850] Step 6:

[0851] Generative AI generates appropriate responses based on the recognized emotions. For example, if the user shows an angry emotion, it generates a calm and polite response. It uses a generative AI model such as GPT-3 to generate responses. The input is the emotion recognition result, and the output is the generated response.

[0852] Step 7:

[0853] The generative AI sends the generated response to the server. The server receives the response and sends it to the terminal. The input is the generated response, and the output is the encrypted response data.

[0854] Step 8:

[0855] The terminal displays the response it receives from the server to the user, for example, "We apologize for the inconvenience. We will check the current status of your order." The input is the encrypted response data, and the output is the text that is displayed to the user.

[0856] (Application example 2)

[0857] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the smart glasses 214 will be referred to as a "terminal".

[0858] Conventional customer service systems using generative AI had problems such as difficulty in ensuring security and responding appropriately to customer emotions. In addition, there was a lack of concrete measures to streamline customer service in brick-and-mortar stores. This could lead to a decrease in customer satisfaction and the risk of information leaks.

[0859] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0860] In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for recognizing the user's emotions using an emotion engine and adjusting the generative AI's response based on the emotions, and a means for supporting customer service in a physical store using a smart device. This makes it possible to provide appropriate responses according to the customer's emotions while ensuring security, and to improve the efficiency of customer service in a physical store.

[0861] The "Secure Proxy SaaS Service for Generative AI" is a cloud-based service that ensures security when using generative AI and provides governance such as blocking specific wording.

[0862] A "server" is a computer system that processes queries to the generative AI and performs security management and data filtering.

[0863] A "means of querying a generative AI via a server" is a method for sending data or questions to a generative AI via a server and receiving a response.

[0864] "Means of server management" are methods for maintaining the operation and security of a server, and for implementing governance such as blocking specific text.

[0865] "Measures to enable governance, such as blocking specific wording" refers to a filtering function to remove confidential information and inappropriate wording from the data processed by the generative AI.

[0866] An "emotion engine" is software that analyzes text input from a user and recognizes the emotions contained in that text.

[0867] A "smart device" is a portable electronic device with Internet connectivity, such as a smartphone or tablet.

[0868] "Means to support customer service in physical stores" refer to technologies and methods for making customer service in physical stores more efficient and improving customer satisfaction.

[0869] This invention is a system for improving customer satisfaction by streamlining customer service in brick-and-mortar stores. This system is composed of a secure proxy SaaS service for generative AI, a server, an emotion engine, and a smart device.

[0870] System Configuration

[0871] 1. Secure proxy SaaS service for generative AI:

[0872] The server provides a secure proxy SaaS service for generative AI. This service ensures security when using generative AI and provides governance such as blocking specific wording.

[0873] 2. Server:

[0874] The server will be installed within the company and will handle queries to the generative AI. The server will be managed using operational and security knowledge cultivated in the information and communications industry.

[0875] 3. Emotion Engine:

[0876] The emotion engine analyzes text input from users, recognizes the emotions contained within that text, and adjusts the generative AI's response based on the recognized emotions.

[0877] 4. Smart Devices:

[0878] Smart devices (such as smartphones and tablets) are used to assist in customer service in brick-and-mortar stores, allowing store staff to respond quickly and appropriately to customer inquiries.

[0879] System Operation

[0880] The server receives customer inquiries from smart devices and uses an emotion engine to recognize the emotion contained in the inquiry. The recognized emotion and the inquiry content are sent to the generative AI, which generates an appropriate response. The generated response is then sent back to the smart device after filtering to block certain phrases.

[0881] Hardware and Software Used

[0882] Hardware: Servers, smartphones, tablets

[0883] Software: Secure proxy SaaS service for generative AI, emotion recognition API, generative AI API

[0884] Examples

[0885] For example, if a customer asks, "Can you tell me more about this product?", the emotion engine will recognize the customer's interest and the generative AI will provide a detailed product explanation. If a customer asks, "I would like to return a product," the emotion engine will recognize the customer's dissatisfaction and the generative AI will provide a polite response.

[0886] Examples of prompt statements

[0887] "Please tell me more about this product"

[0888] "I'd like to return it."

[0889] "Do you have this item in stock?"

[0890] In this way, a system for supporting customer service in brick-and-mortar stores can be realized. This system provides appropriate responses according to the customer's emotions while ensuring security, making it possible to improve the efficiency of customer service in brick-and-mortar stores.

[0891] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0892] Step 1:

[0893] A user enters a query using a smart device.

[0894] Input: User's query text (e.g. "Tell me more about this product")

[0895] Output: The query text is sent from the smart device to the server.

[0896] Step 2:

[0897] The server sends the received query text to an emotion engine to recognize emotions.

[0898] Input: Query text

[0899] Output: Perceived emotion (e.g. interest, anger, frustration)

[0900] Specific operation: The server calls the emotion recognition API and analyzes the query text to recognize emotions.

[0901] Step 3:

[0902] The server sends the recognized emotion and the query text to a generative AI to generate an appropriate response.

[0903] Input: query text, recognized sentiment

[0904] Output: The generated response text

[0905] How it works: The server calls the generative AI API, sends the query text and the recognized emotion as a prompt, and the generative AI generates a response.

[0906] Step 4:

[0907] The server filters the generated response text to block certain phrases.

[0908] Input: Generated response text

[0909] Output: The filtered response text

[0910] What happens: The server applies a filtering algorithm to filter out sensitive information and inappropriate language.

[0911] Step 5:

[0912] The server sends the filtered response text to the smart device.

[0913] Input: Filtered response text

[0914] Output: Response text displayed on smart device

[0915] Specific operation: The server sends the filtered response text to the smart device and displays it to the user.

[0916] Step 6:

[0917] The user checks the response on their smart device and makes additional enquiries if necessary.

[0918] Input: Filtered response text

[0919] Output: Additional query text from the user (e.g. "What other products do you recommend?")

[0920] Specific Actions: The user checks the response on the smart device and enters additional inquiries.

[0921] In this way, a system that supports customer service in brick-and-mortar stores is realized. The specific operations at each step ensure security and make it possible to provide appropriate responses according to the customer's emotions.

[0922] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires a voice indicating a user input for the result of the specific processing. The control unit 46A transmits the voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[0923] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by making a neural network perform deep learning. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating a voice, text data indicating a text, and image data indicating an image is input. The data generation model 58 performs inference on the input inference data according to the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0924] Another example of generative AI is Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) are mentioned.

[0925] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0926] [Third embodiment]

[0927] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0928] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0929] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a wide area network (WAN) and / or a local area network (LAN).

[0930] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0931] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs the voice according to instructions from the processor 46.

[0932] Camera 42 is a small digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (e.g., an imaging range defined by an angle of view equivalent to the width of the field of vision of an average healthy person).

[0933] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54. The transmission and reception of various types of information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is performed in a secure state.

[0934] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0935] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32, and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0936] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0937] In the headset type terminal 314, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0938] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0939] "Example 1"

[0940] The present invention is a system that provides a secure proxy SaaS service for generative AI. Specifically, a server is installed within the company, and queries are made to the generative AI via this server. Server management utilizes operational and security knowledge cultivated in the information and communications industry, enabling governance such as blocking specific wording through settings for each individual company.

[0941] "Example 2"

[0942] As a concrete example, if a company wants to use generative AI to respond to customers, they can use the system of the present invention to make inquiries to the generative AI via a server installed within the company. In this case, server management can utilize the operational and security knowledge cultivated in information and communications to set up settings to block specific text, such as customer information or confidential information. This makes it possible to use generative AI in business while ensuring security.

[0943] The process flow of each embodiment will be described below.

[0944] "Example 1"

[0945] Step 1: Set up a server within the company. This server acts as an intermediary for communication with the generative AI.

[0946] Step 2: A query is sent to the generative AI via the server. This query is based on the company's business needs, for example, generating an automatic response for customer support.

[0947] Step 3: Manage the server. This management utilizes operational and security knowledge cultivated in the information and communications industry, enabling governance such as blocking specific wording with settings for each individual company.

[0948] "Example 2"

[0949] Step 1: If a company wants to use generative AI to handle customer support, it first sets up a server within the company.

[0950] Step 2: Make an inquiry to the generative AI via the installed server. This inquiry may involve generating an automatic response to a customer inquiry.

[0951] Step 3: Manage the server. This management utilizes operational and security knowledge cultivated in the information and communications industry, and sets up settings to block specific text, such as customer information or confidential information. This makes it possible to use generative AI in business while ensuring security.

[0952] Example 1

[0953] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal".

[0954] In order to use generative AI models safely and efficiently within a company, it is necessary to minimize the risk of data leakage to the outside while implementing governance based on the company's security policy. However, conventional systems lack the ability to block or filter specific text, making it difficult to fully comply with a company's security policy.

[0955] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0956] In this invention, the server includes a means for providing a secure proxy service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving a prompt text from a user's terminal, a means for inspecting the received prompt text based on filtering rules, a means for relaying the prompt text that has passed the inspection to the generative AI model, a means for receiving a response from the generative AI model and inspecting it again based on filtering rules, and a means for transmitting the response that has passed the filtering to the user's terminal. This makes it possible to safely and efficiently use the generative AI model while realizing governance based on the company's security policy.

[0957] The "Secure Proxy Service for Generative AI" is a service that securely relays queries to generative AI models and enables governance based on a company's security policies.

[0958] A “server” is a computer system installed within a company that relays queries to the generative AI model and provides filtering and governance functions.

[0959] A "prompt" is text that contains instructions or questions that a user enters into a generative AI model.

[0960] "Filtering rules" are rules or conditions for blocking specific words or phrases, and are set based on a company's security policy.

[0961] A "generative AI model" is an artificial intelligence model that generates responses based on prompts from the user.

[0962] "Governance" refers to management and control, such as blocking specific words or phrases, based on a company's security policy.

[0963] A "user's device" is a device, such as a computer or smartphone, that a user uses to input prompts to a generative AI model and receive responses.

[0964] An "inquiry" refers to instructions or questions that a user gives to a generative AI model, and is sent to the server as a prompt.

[0965] A "response" is an answer or piece of information that a generative AI model generates based on a prompt.

[0966] This invention is a system that provides a secure proxy service for generative AI, and queries generative AI models through a server installed within a company. A specific embodiment of this system will be described below.

[0967] Server installation and initial configuration

[0968] The server is installed within the company and connected to the company's network. The server administrator accesses the server's settings screen and sets filtering rules based on the company's security policy. For example, a rule is set to block inquiries that include the phrase "confidential information." Once the settings are complete, the server is connected to the company's network and begins operation.

[0969] User query input

[0970] The user inputs a prompt for the generative AI model from their own device. For example, a marketing person inputs a prompt such as, "Please tell me your sales forecast for the next quarter." Once the input is complete, the user clicks the send button.

[0971] The server receives and examines the query

[0972] The server receives the prompt text sent from the user's terminal. It checks the received prompt text based on the filtering rules. For example, it checks whether the text contains the phrase "confidential information." If there are no problems, it proceeds to the next step.

[0973] Relaying inquiries

[0974] The server relays the prompts that pass inspection to the generative AI model, which generates a response based on the prompts received. For example, it generates a response like, "Sales forecast for the next quarter is XXX yen."

[0975] Receiving and filtering responses by the server

[0976] The server receives the response from the generative AI model. The server checks the response again based on the filtering rules. For example, it checks whether the response contains the phrase "confidential information." If there are no problems, it proceeds to the next step.

[0977] Providing a response

[0978] The server sends the filtered responses to the user's device. The user can check the response from the generative AI model on the device. For example, the user can check the response, "Sales forecast for the next quarter is XXX yen."

[0979] Examples

[0980] As a concrete example, Company A has a setting to block queries that contain the phrase "confidential information." When a user enters the prompt phrase "Please generate a report that contains confidential information," the server detects this phrase and returns an error message without relaying the query.

[0981] Company B also uses a generative AI model to generate marketing reports. When a user enters a prompt such as "Please generate a report on the latest market trends," the server inspects the query and relays it to the generative AI model if it determines there is no problem. The response from the generative AI model is provided to the user via the server.

[0982] The system enables companies to use generative AI models safely and efficiently.

[0983] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0984] Step 1:

[0985] The server is installed within the company and connected to the network. The server administrator accesses the server's configuration screen and sets filtering rules based on the company's security policy. For example, a rule is set to block inquiries that include the phrase "confidential information." The input is the security policy and filtering rules, and the output is the configured server.

[0986] Step 2:

[0987] The user inputs a prompt sentence to the generative AI model from his / her own device. For example, the user inputs a prompt sentence such as "Please tell me your sales forecast for the next quarter." The input is the prompt sentence, and the output is the sent prompt sentence. The user clicks the send button.

[0988] Step 3:

[0989] The server receives the prompt text sent from the user's terminal. It checks the received prompt text based on filtering rules. For example, it checks whether it contains the phrase "confidential information." The input is the sent prompt text, and the output is the inspection result. If there are no problems, it proceeds to the next step.

[0990] Step 4:

[0991] The server relays the prompts that pass inspection to the generative AI model. The generative AI model generates a response based on the prompts it receives. For example, it generates a response like, "Sales forecast for the next quarter is XXX yen." The input is the prompt that passes inspection, and the output is the generated response.

[0992] Step 5:

[0993] The server receives the response from the generative AI model. The received response is inspected again based on the filtering rules. For example, it checks whether it contains the phrase "confidential information." The input is the response from the generative AI model, and the output is the inspection result. If there are no problems, it proceeds to the next step.

[0994] Step 6:

[0995] The server sends the filtered response to the user's device. The user can check the response from the generative AI model on the device. For example, the user can check the response "Sales forecast for the next quarter is XXX yen." The input is the filtered response, and the output is the response sent to the user's device.

[0996] (Application example 1)

[0997] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[0998] When using generative AI within a company, there is a risk that confidential information may be leaked to the outside. In addition, it is difficult to implement appropriate governance when specific wording or data is included. Furthermore, security must be ensured when making inquiries to generative AI using smartphones.

[0999] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means. In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing a server, a means for making an inquiry to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for making an inquiry to the generative AI through an application installed on a smartphone, a means for securely communicating via a secure proxy server, a means for sending a filtered prompt to the generative AI, and a means for displaying the result on a smartphone application. This makes it possible to reduce the risk of confidential information leakage and perform appropriate governance when using generative AI within a company. In addition, high security can be ensured even when making inquiries to the generative AI using a smartphone.

[1000] The "Secure Proxy SaaS Service for Generative AI" is a cloud-based service that provides a proxy server for securely making inquiries to generative AI.

[1001] A "server" is a computer system that relays queries to the generative AI and manages security and governance.

[1002] A "means for querying a generative AI via a server" is a method for sending a prompt to a generative AI via a server and receiving a result.

[1003] "Means of server management" refers to the methods for maintaining and properly managing the operation and security of the server.

[1004] "Measures that enable governance, such as blocking specific wording" refers to methods for filtering out specific keywords and phrases and controlling the content of queries made to generative AI.

[1005] An "application installed on a smartphone" is software that runs on a smartphone and is used to make queries to the generative AI.

[1006] "Means for securely communicating via a secure proxy server" refers to a method for encrypting data and communicating securely through a secure proxy server.

[1007] "Means for transmitting a filtered prompt to the generative AI" refers to a method for transmitting the prompt after it has been filtered to the generative AI.

[1008] "Means for displaying results in a smartphone application" refers to a method for displaying the results returned from the generative AI in an application on a smartphone.

[1009] The following system configuration will be described as an embodiment of the present invention.

[1010] System Configuration

[1011] Hardware

[1012] Server: A computer system installed within a company that provides a secure proxy SaaS service for generative AI.

[1013] Smartphone: A device that allows users to make inquiries to generative AI.

[1014] software

[1015] Secure Proxy Server: Runs on the server, relays queries to the generative AI, and manages security and governance.

[1016] Smartphone application: Software that is installed on a smartphone and allows for queries to be made to the generative AI.

[1017] Data processing and calculation

[1018] server

[1019] The server provides a secure proxy SaaS service for generative AI. When a user makes a query to generative AI through a smartphone application, the prompt is first sent to the server. The server applies governance such as blocking certain words, and sends the filtered prompt to the generative AI.

[1020] Smartphone

[1021] A user uses a smartphone application to make a query to the generative AI. After entering a prompt and pressing the send button, the prompt is sent to the generative AI via a secure proxy server. The results returned from the generative AI are then displayed on the smartphone application via the secure proxy server again.

[1022] Examples

[1023] For example, when a company queries a generative AI for documents containing confidential information, the system can be used to reduce the risk of confidential information leaking to the outside. The user enters the following prompt into a smartphone application:

[1024] Prompt: "Generate the following sentence: Details about a company's new product."

[1025] This prompt is sent to a server where it is filtered to see if it contains specific words, and the filtered prompt is sent to a generative AI, which then displays the generated results on a smartphone application.

[1026] In this way, it is possible to reduce the risk of confidential information leaks and implement appropriate governance when using generative AI within a company. It is also possible to ensure high security when making inquiries to generative AI using smartphones.

[1027] The flow of the specific process in the application example 1 will be described with reference to FIG.

[1028] Step 1:

[1029] The user starts the smartphone application and inputs the question (prompt) they want to ask the generative AI. The input prompt is temporarily saved in the application.

[1030] Step 2:

[1031] When the user presses the send button, the smartphone application sends the prompt text to the secure proxy server. The input is the prompt text, and the output is a send request to the secure proxy server.

[1032] Step 3:

[1033] The server analyzes the received prompt text and filters it to see if it contains specific words. The input is the prompt text, and the output is the filtered prompt text. Specifically, the server checks it against a specific keyword list and removes any inappropriate words that it finds.

[1034] Step 4:

[1035] The server sends the filtered prompt to the generative AI. The input is the filtered prompt, and the output is a request to send to the generative AI. Specifically, the server sends an HTTP request to the API endpoint of the generative AI.

[1036] Step 5:

[1037] The generative AI generates a response based on the received prompt and sends the result back to the server. The input is the filtered prompt and the output is the generated response. In concrete terms, the generative AI uses its internal model to generate a response to the prompt.

[1038] Step 6:

[1039] The server receives the response returned from the generative AI and sends it to the smartphone application. The input is the generated response, and the output is a request to send to the smartphone application. In concrete terms, the server sends an HTTP request with the response data to the API endpoint of the smartphone application.

[1040] Step 7:

[1041] The smartphone application displays the response received from the server. The input is the generated response and the output is the result displayed to the user. In concrete terms, the application passes the received data to a UI component for display on the screen.

[1042] Example 2

[1043] Next, a second embodiment of the second form will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal".

[1044] Customer support systems that use generative AI are required to respond to customers efficiently while ensuring security. However, with conventional systems, there is a risk that customer information or confidential information may be sent to the generative AI model, which may lead to security issues. In addition, responses from the generative AI model may be difficult for users to understand, which can lead to a decline in the quality of customer support. To solve these issues, a system that can filter the content of inquiries and convert the content of responses is needed.

[1045] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1046] In this invention, the server includes a means for receiving an inquiry from a terminal, a means for analyzing the received inquiry content and filtering out specific wording, a means for generating a prompt sentence based on the filtered inquiry content, a means for sending the generated prompt sentence to a generative AI model, a means for receiving a response from the generative AI model, a means for analyzing the received response and converting it into a format that is easy for a user to understand, a means for sending the converted response to the terminal, a means for installing a server, a means for making an inquiry to the generative AI model via the server, a means for managing the server, and a means for enabling governance such as blocking specific wording. This enables efficient customer service using generative AI while ensuring security.

[1047] A "terminal" is a device through which a user inputs inquiry content and transmits it to a server.

[1048] A "server" is a device that receives queries from a terminal, queries the generative AI model, and sends responses to the terminal.

[1049] "Inquiry content" is text data of a question or request that a user inputs via a terminal.

[1050] "Filtering" is the process of removing or masking specific wording (such as customer information or confidential information) from the content of an inquiry.

[1051] A "prompt sentence" is the text of a query that is generated to be sent to a generative AI model.

[1052] A "generative AI model" is an artificial intelligence model that generates a response based on an input prompt sentence.

[1053] A "response" is the text data of the answer returned from the generative AI model.

[1054] "Analysis" is the process of understanding the inquiry and response, and extracting or converting the necessary information.

[1055] "Transformation" is the process of changing the response from a generative AI model into a form that is easy for a user to understand.

[1056] "Governance" refers to the means of managing the operation and security of a system, and exercising control such as blocking specific text.

[1057] This invention relates to a customer service system that uses generative AI. Specifically, it is a system that uses a terminal, a server, and a generative AI model to filter inquiries and convert responses.

[1058] Hardware and Software Used

[1059] The terminal is a device where a user inputs the contents of an inquiry and transmits the input to the server. The terminal may be a PC, a smartphone, a tablet, or the like.

[1060] The server is a device that receives queries from the terminal, queries the generative AI model, and sends the response to the terminal. The server has installed software for parsing the query content, filtering, generating prompts, and parsing and converting the response. Specific software includes a natural language processing (NLP) library and an API client.

[1061] A generative AI model is an artificial intelligence model that generates responses based on input prompt sentences. Specifically, OpenAI's GPT-4 and other models are used.

[1062] Data processing and calculation

[1063] The terminal transmits the inquiry content input by the user to the server. For example, the user inputs "Please tell me the price of the new product."

[1064] The server analyzes the received inquiry content and filters out specific phrases (such as customer information or confidential information). Based on the filtered inquiry content, it generates a prompt sentence suitable for the generative AI model. For example, it generates a prompt sentence such as "Please tell me the price of the new product."

[1065] The server sends the generated prompt sentence to the generative AI model, which generates a response based on the prompt sentence and sends it back to the server. For example, it generates a response such as "The price of the new product is $100."

[1066] The server parses the response and converts it into a format that is easy for the user to understand. It then sends the converted response to the terminal. The terminal displays the response to the user. For example, the message "The price of the new product is $100" is displayed on the user's screen.

[1067] Examples

[1068] For example, if a company receives an inquiry from a customer saying, "Please tell me the price of a new product," the following process will take place:

[1069] 1. On the terminal, the user enters "Please tell me the price of the new product" and clicks the send button.

[1070] 2. The server analyzes the received data and extracts the text "Please tell me the price of the new product." It then performs a filtering process to check whether the data contains a specific phrase.

[1071] 3. The server generates a prompt sentence based on the filtered text: "Please tell me the price of the new product."

[1072] 4. The server creates and sends an API request to send the generated prompt to GPT-4.

[1073] 5. The server receives the response from GPT-4 and gets the text “The price of the new product is $100.”

[1074] 6. The server parses the received text and replaces technical terms with more common language where necessary.

[1075] 7. The server sends the converted response to the terminal, which displays the received response to the user, and the message "The new product costs $100" appears on the user's screen.

[1076] In this way, businesses can leverage generative AI to deliver efficient customer service while ensuring security.

[1077] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1078] Step 1:

[1079] The terminal transmits the inquiry content entered by the user to the server. As input, the user inputs "Please tell me the price of the new product" and clicks the send button. As output, the inquiry content is transmitted to the server.

[1080] Step 2:

[1081] The server analyzes the inquiry received from the terminal. As input, it receives the text data "Please tell me the price of the new product" sent from the terminal. To process the data, it analyzes the text using natural language processing (NLP) technology and filters out specific phrases (such as customer information or confidential information). As output, it obtains the filtered inquiry content.

[1082] Step 3:

[1083] The server generates a prompt sentence based on the filtered inquiry content. As input, it receives the filtered text "Please tell me the price of the new product." As data processing, it converts it into a format suitable for the generative AI model. As output, it obtains the generated prompt sentence.

[1084] Step 4:

[1085] The server sends the generated prompt sentence to the generative AI model. As input, it receives the generated prompt sentence "Please tell me the price of the new product." As a specific operation, it creates an API request and sends it to the generative AI model (e.g., GPT-4). As output, the prompt sentence is sent to the generative AI model.

[1086] Step 5:

[1087] The server receives the response from the generative AI model. As input, it receives the response data returned from the generative AI model. As a specific operation, it parses the API response and obtains the text "The price of the new product is $100." As output, it obtains the response from the generative AI model.

[1088] Step 6:

[1089] The server analyzes the received response and converts it into a format that is easy for the user to understand. As input, it receives the response data "The price of the new product is $100." As data processing, it performs conversion processing such as replacing technical terms with common terms. As output, it obtains the converted response.

[1090] Step 7:

[1091] The server sends the converted response to the terminal. As input, it receives the converted response data "The price of the new product is $100." As a specific operation, it sends the response data to the terminal. As output, the response is sent to the terminal.

[1092] Step 8:

[1093] The terminal displays the received response to the user. As input, it receives the response data "The price of the new product is $100" sent from the server. As a specific operation, it displays the response message on the user's screen. As output, the user can confirm the response.

[1094] (Application example 2)

[1095] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[1096] Conventional customer service systems using generative AI have difficulty in ensuring security and cannot block certain text (e.g., customer information or confidential information), making it difficult to use safely within a company. In addition, they lacked the functionality to make personalized product recommendations to users, making it difficult to improve customer satisfaction.

[1097] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1098] In this invention, the server includes a means for providing a secure proxy service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving questions from users and sending them to the generative AI, a means for returning answers from the generative AI to the user, and a means for making personalized product recommendations based on the user's purchase history and browsing history. This enables safe customer service and personalized product recommendations using generative AI while ensuring security.

[1099] The "Secure Proxy Service for Generative AI" is a relay service that enables safe inquiries to be made to generative AI, and exchanges data while ensuring security.

[1100] A "server" is a computer system that stores, processes, and manages data on a network.

[1101] A "means for querying a generative AI via a server" is a method for sending questions or instructions to a generative AI via a server.

[1102] "Server management" refers to the activity of operating, maintaining, and managing the security of a server.

[1103] "Governance such as blocking specific wording" is a management technique that filters out specific keywords and phrases (such as customer information or confidential information) to prevent the leakage of inappropriate information within the system.

[1104] "Means for receiving questions from a user and sending them to a generative AI" refers to a method for receiving questions entered by a user and sending them to a generative AI.

[1105] "Means of returning an answer from a generative AI to a user" refers to a method for returning an answer obtained from a generative AI to a user.

[1106] "Means for personalized product recommendations based on a user's purchase history and browsing history" refers to a method for analyzing a user's past purchase history and browsing history and recommending individually optimized products based on that.

[1107] A system for implementing this invention includes a server that provides a secure proxy service for the generative AI, a terminal that receives questions from users and sends them to the generative AI, a terminal that returns answers from the generative AI to the user, and a function that makes personalized product recommendations based on the user's purchasing history and browsing history.

[1108] Hardware and Software Configuration

[1109] Server: Installed within a company, it provides a secure proxy service for generative AI. The server is built using a web framework such as Flask.

[1110] Device: A device used by a user, such as a smartphone or tablet. These devices run applications that receive questions from users and send them to the generative AI.

[1111] Generative AI: Use generative AI models such as OpenAI's GPT-3. Generative AI generates appropriate answers to questions posed by users.

[1112] Data processing and calculation

[1113] 1. Receiving user questions:

[1114] The user inputs a question through the terminal, for example, "How do you use this product?"

[1115] 2. Filter questions:

[1116] The server filters certain phrases (customer information and confidential information) from the question, ensuring security.

[1117] 3. Querying the generative AI:

[1118] The filtered questions are sent to a generative AI, which generates answers to the questions.

[1119] 4. Response to the answer:

[1120] The server receives the answer from the generative AI and responds to the user through the device, for example, "This product is used in the following steps..."

[1121] 5. Personalized product recommendations:

[1122] The server analyzes the user's purchase and browsing history and recommends individually optimized products based on that information, for example by displaying "These are the products we recommend for you."

[1123] Examples

[1124] When a user asks, "How do you use this product?", the server filters the question and sends it to the generative AI. The generative AI responds with, "This product is used in the following steps..." and the answer is returned to the user. Related products are also recommended based on the user's past purchase and browsing history.

[1125] Examples of prompt statements

[1126] "How do you use this product?"

[1127] In this way, secure customer interactions and personalized product recommendations using generative AI can be achieved while ensuring security.

[1128] The flow of the specific process in the application example 2 will be described with reference to FIG.

[1129] Step 1:

[1130] A user inputs a question through a terminal, for example, "How do you use this product?" The input question is sent to a server by an application on the terminal.

[1131] Step 2:

[1132] The server filters the questions it receives. Specifically, it detects specific phrases (customer information or confidential information) from the questions and blocks or masks them. The filtered questions are then sent to the generative AI.

[1133] Step 3:

[1134] The server sends the filtered questions to the generative AI, which generates appropriate answers to the questions it receives. For example, it uses OpenAI's GPT-3 to generate an answer like, "This product is used in the following steps..."

[1135] Step 4:

[1136] The server receives the answer from the generative AI. The server returns the answer to the device. The device displays the answer from the generative AI to the user.

[1137] Step 5:

[1138] The server analyzes the user's purchase history and browsing history. Specifically, it retrieves the user's past purchase history and browsing history from the database and makes personalized product recommendations based on that. For example, it displays "These are the products we recommend for you."

[1139] Step 6:

[1140] The server sends the personalized product recommendations to the terminal. The terminal displays the recommended products to the user. The user can review the recommended products and consider purchasing them.

[1141] In this way, secure customer interactions and personalized product recommendations using generative AI can be achieved while ensuring security.

[1142] In addition, an emotion engine that estimates the emotion of the user may be further combined. That is, the identification processing unit 290 may estimate the emotion of the user using the emotion identification model 59, and perform identification processing using the emotion of the user.

[1143] "Example 1"

[1144] As one embodiment of the present invention, there is a system that provides a secure proxy SaaS service for generative AI, which is combined with an emotion engine. This system queries the generative AI through a server installed in the company and manages the server. The server management utilizes operational and security knowledge cultivated in information and communications, enabling governance such as blocking specific wording. Furthermore, the generative AI includes an emotion engine that recognizes the user's emotions. Specifically, it receives text input from the user and recognizes the user's emotions from the text. Then, the generative AI's response is adjusted based on the recognized emotion. For example, if a user inputs text indicating an emotion of anger, the emotion engine recognizes the anger, and the generative AI generates an appropriate response based on the recognition result. This enables a more appropriate response that corresponds to the user's emotions.

[1145] "Example 2"

[1146] As one embodiment of the present invention, there is a system that provides a secure proxy SaaS service for generative AI, which is combined with an emotion engine. This system queries the generative AI through a server installed in the company and manages the server. The server management utilizes operational and security knowledge cultivated in information and communications, enabling governance such as blocking specific wording. Furthermore, the generative AI includes an emotion engine that recognizes the user's emotions. Specifically, it receives text input from the user and recognizes the user's emotions from the text. Then, the generative AI's response is adjusted based on the recognized emotion. For example, if a user inputs text indicating an emotion of anger, the emotion engine recognizes the anger, and the generative AI generates an appropriate response based on the recognition result. This enables a more appropriate response that corresponds to the user's emotions.

[1147] The process flow of each embodiment will be described below.

[1148] "Example 1"

[1149] Step 1: Receive text input from a user, which may include the user's sentiment.

[1150] Step 2: Input the received text into the emotion engine, which recognizes the user's emotion from the text.

[1151] Step 3: The output from the emotion engine (the recognized emotion) is fed into the generative AI, which generates a response based on this emotion.

[1152] Step 4: The generative AI's response is returned to the user. This response corresponds to the user's emotion.

[1153] Example 1

[1154] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal".

[1155] Conventional services using generative AI have security and governance issues, making it difficult to block specific text or operate based on a company's security policy. In addition, they are unable to generate responses that take into account the user's emotions, which can lead to a poor user experience.

[1156] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1157] In this invention, the server includes a means for providing a secure proxy service for the generative AI, a means for installing the server, a means for making an inquiry to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific words, a means for including an emotion engine for recognizing the user's emotions, and a means for adjusting the response of the generative AI based on the emotions recognized by the emotion engine. This makes it possible to generate an appropriate response corresponding to the user's emotions while strengthening security and governance.

[1158] "Generative artificial intelligence" is an artificial intelligence technology that generates natural language responses and content based on user input.

[1159] A "secure proxy service" is a service that relays user requests and provides proxy functions to enhance security and governance.

[1160] A "server" is a computer system that processes and manages data on a network.

[1161] "Governance" refers to the management and control of system operations, such as blocking specific wording and applying security policies.

[1162] An "emotion engine" is software or hardware for analyzing text input from a user and recognizing its emotions.

[1163] "Adjusting the response" means changing the content of the response generated by the generative artificial intelligence based on the emotion recognized by the emotion engine.

[1164] The present invention is a system that provides a secure proxy service for querying a generative artificial intelligence through a server installed within a company. The system includes an emotion engine that recognizes the user's emotions and adjusts responses based on the emotions.

[1165] Hardware and Software Used

[1166] 1. Server

[1167] It is installed within a company.

[1168] It will be managed utilizing operational and security knowledge cultivated in the information and communications industry.

[1169] It has a governance function that blocks specific wording.

[1170] 2. Terminal

[1171] A device (e.g. PC, smartphone) that allows a user to input text.

[1172] 3. Generative Artificial Intelligence

[1173] An artificial intelligence technology that generates natural language responses based on user input.

[1174] 4. Emotion Engine

[1175] Software that analyzes text input from users and recognizes their emotions.

[1176] Data processing and calculation

[1177] 1. User text input

[1178] A user uses the terminal to input text.

[1179] For example: "Why is it so slow?"

[1180] 2. Receiving and parsing the text by the server

[1181] The server receives text input from the user and checks whether it contains specific words.

[1182] Example: Detect the keyword "slow".

[1183] 3. Emotion Recognition by Emotion Engine

[1184] The server sends the text to an emotion engine, which analyzes the text to recognize the user's emotion.

[1185] Example: The emotion engine recognizes "anger."

[1186] 4. Response generation using generative artificial intelligence

[1187] The server sends the emotion recognition results to the generative artificial intelligence, which then generates an appropriate response.

[1188] For example: "We apologize for the inconvenience. We are currently working on improving our system."

[1189] 5. Server Sends Response

[1190] The server sends the generated response to the user's terminal.

[1191] Examples

[1192] For example, if a user enters the text "Why is it so slow?", the following happens:

[1193] 1. User text input

[1194] A user types in the chat window, "Why is this so slow?"

[1195] 2. Receiving and parsing the text by the server

[1196] The server receives the text "Why is it so slow?" and detects the keyword "slow".

[1197] 3. Emotion Recognition by Emotion Engine

[1198] The server sends the text to an emotion engine, which recognizes "anger."

[1199] 4. Response generation using generative artificial intelligence

[1200] The server sends the emotion recognition results to the generative artificial intelligence, which generates a response saying, "We apologize for the inconvenience. We are currently working on improving the system."

[1201] 5. Server Sends Response

[1202] The server displays the generated response in the user's chat window.

[1203] Examples of prompt statements

[1204] Here are some example prompts to input to a generative AI model:

[1205] User: Why is it so slow?

[1206] Emotion engine: Anger recognition

[1207] Generative AI: We apologize for the inconvenience. We are currently working on improving the system.

[1208] In this way, a more appropriate response can be made according to the user's emotions.

[1209] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1210] Step 1:

[1211] Initial Server Setup

[1212] The server is installed within the company. The server administrator accesses the server's settings screen and sets filtering rules and security policies. Specifically, they set rules to block specific wording and apply security policies. The input is the administrator's settings information, and the output is the state of the server with the settings reflected.

[1213] Step 2:

[1214] User text input

[1215] A user uses a terminal to input text, specifically, typing "why is this so slow?" into a chat window. The input is the user's text, and the output is the transmission of text data from the terminal to the server.

[1216] Step 3:

[1217] Receiving and parsing text by the server

[1218] The server receives text input from the user. It analyzes the received text and checks whether it contains a specific phrase. Specifically, it detects the keyword "slow." The input is the user's text data, and the output is the analysis result.

[1219] Step 4:

[1220] Emotion recognition by emotion engine

[1221] The server sends the text to the emotion engine. The emotion engine analyzes the text and recognizes the user's emotion. Specifically, the emotion engine recognizes "anger." The input is the analyzed text data, and the output is the emotion recognition result.

[1222] Step 5:

[1223] Response generation using generative artificial intelligence

[1224] The server sends the emotion recognition results to the generative AI. The generative AI generates an appropriate response based on the emotion recognition results. Specifically, it generates a response such as "We apologize for the inconvenience. We are currently working on improving the system." The input is the emotion recognition results, and the output is the generated response.

[1225] Step 6:

[1226] Server sends response

[1227] The server sends the generated response to the user's terminal, and displays the generated response in a chat window. The input is the generated response data, and the output is the response displayed on the user's terminal.

[1228] (Application example 1)

[1229] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[1230] Conventional generative AI systems were unable to respond in a way that took the user's emotions into account, and had insufficient functions for blocking specific phrases and for governance. This resulted in security risks and a poor user experience when used within a company. It was also difficult to provide appropriate responses when used on mobile devices such as smartphones.

[1231] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1232] In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific words, a means for recognizing the user's emotions using an emotion engine and adjusting the generative AI's response based on the emotion, and a means for functioning as an application installed on a smartphone. This makes it possible to reduce security risks within a company and provide appropriate responses according to the user's emotions.

[1233] The "Secure Proxy SaaS Service for Generative AI" is a Software as a Service (SaaS) that provides a proxy service for securely making inquiries to generative artificial intelligence (AI).

[1234] "Server" refers to a computer system that runs the secure proxy SaaS service for generative AI and processes and manages data within the company.

[1235] A "means of querying a generative AI via a server" is a method for sending data or requests to a generative AI via a server and receiving a response.

[1236] "Means of server management" refers to methods for managing the operation and security of servers and providing stable services.

[1237] "Measures to enable governance, such as blocking specific wording" refers to a method that provides the ability to detect and block specific keywords or phrases based on a company's security policy.

[1238] An "emotion engine" is a software component for analyzing emotions from a user's input text and generating an appropriate response based on the emotions.

[1239] "Means to function as an application installed on a smartphone" refers to software that operates on a smartphone and is provided in a form that the user can use.

[1240] The system for implementing this invention is composed mainly of a server that provides a secure proxy SaaS service for generative AI. The server is installed within a company and provides a proxy service for securely making inquiries to generative AI. The server utilizes operational and security knowledge cultivated in information and communications to enable governance such as blocking specific wording.

[1241] The server is equipped with an emotion engine that recognizes the user's emotions, receives text input from the user, analyzes the emotions from the text, and uses the TextBlob library to obtain the polarity of the emotions and adjusts the generative AI's response based on the user's emotions.

[1242] Users access the system through an application installed on their smartphones, which sends user input to the server and receives responses from it, which checks for specific phrases, generates prompts based on sentiment analysis, and uses the OpenAI API to generate appropriate responses.

[1243] For example, if a user enters "I'm very tired today," the emotion engine recognizes that the user is tired (a negative emotion). The server generates a prompt sentence like this:

[1244] "The user is tired. Generate an appropriate response to the following message: I'm very tired today."

[1245] Based on this prompt, a response is generated using the OpenAI API, and provided to the user, for example, "Thank you for your hard work. Please take it easy and get some rest today."

[1246] This system can reduce security risks within a company and provide appropriate responses according to the user's emotions.

[1247] The flow of the specific process in the application example 1 will be described with reference to FIG.

[1248] Step 1:

[1249] A user starts an application installed on a smartphone and inputs text. The input text is a sentence in a natural language that includes the user's emotions and intentions. The input data is sent from the user's smartphone to a server.

[1250] Step 2:

[1251] The server parses the received text input and checks for specific terms. It checks whether the text contains any of the keywords or phrases in the list of specific terms. If the text contains any of the specific terms, the server generates a warning message and sends it back to the user. The input data is the text input, and the output data is either a warning message or text input for the next step.

[1252] Step 3:

[1253] If no specific phrase is included, the server performs sentiment analysis on the text input using the sentiment engine. It uses the TextBlob library to calculate the sentiment polarity (positive, negative, neutral) of the text. The input data is the text input, and the output data is the sentiment polarity.

[1254] Step 4:

[1255] The server generates a prompt sentence based on the result of the sentiment analysis. For example, if the sentiment polarity is negative, the server creates a prompt sentence in the form of "The user is tired. Please generate an appropriate response to the following message: [Text input]". The input data is the sentiment polarity and the text input, and the output data is the prompt sentence.

[1256] Step 5:

[1257] The server sends the generated prompt to the OpenAI API, which generates an appropriate response. The OpenAI API generates a natural language response based on the prompt and sends it back to the server. The input data is the prompt, and the output data is the generated response.

[1258] Step 6:

[1259] The server sends the generated response to the user's smartphone. The user receives and displays the response through an application on the smartphone. The input data is the generated response, and the output data is the response displayed on the user's smartphone.

[1260] Through the above processing steps, it is possible to provide an appropriate response according to the user's emotions.

[1261] Example 2

[1262] Next, a second embodiment of the second form will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal".

[1263] Conventional systems using generative AI were unable to respond in a way that took the user's emotions into account, and there were also security concerns. In particular, when used within a company, there was a high risk of customer information and confidential information being leaked, and appropriate governance was required to prevent this. In addition, because it was not possible to generate appropriate responses based on the user's emotions, it was difficult to improve customer satisfaction.

[1264] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1265] In this invention, the server includes a means for providing a secure proxy service for the generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving text input from a user and sending the text to the generative AI, a means for the generative AI to recognize the user's emotions and generate a response based on the emotions, and a means for returning the generated response to the user. This makes it possible to generate an appropriate response corresponding to the user's emotions while ensuring security.

[1266] The "Secure Proxy Service for Generative AI" is a relay service that allows secure access to generative AI and has the ability to filter and encrypt data.

[1267] A "server" is a computer system that provides services to other computers over a network.

[1268] "Means of querying the generative AI via a server" refers to a function for sending data to the generative AI via a server and receiving the response.

[1269] "Server management" refers to the set of activities for operating, monitoring, and maintaining a server, including security measures and performance optimization.

[1270] "Governance such as blocking specific wording" is a management technique for preventing information leaks by detecting specific keywords or phrases and blocking or masking them.

[1271] "Means for receiving text input from a user and sending that text to the generative AI" refers to a function for receiving text data entered by a user and sending it to the generative AI.

[1272] "Means for the generative AI to recognize the user's emotions and generate a response based on those emotions" refers to the function that enables the generative AI to analyze emotions from the user's input text and generate an appropriate response based on the results of that analysis.

[1273] "Means of returning the generated response to the user" refers to a function for sending the response generated by the generative AI to the user via a server and displaying it to the user.

[1274] The present invention relates to a system for providing a secure proxy service for generative AI. A specific embodiment of this system will be described below.

[1275] System Configuration

[1276] server

[1277] The server is installed within the company and manages inquiries to the generative AI. The server utilizes operational and security knowledge cultivated in the information and communications industry to set up the system to block specific text (e.g., customer information or confidential information). The server uses the HTTPS protocol to encrypt data and ensure security.

[1278] Terminal

[1279] The terminal receives text input from the user and transmits the text to the server. The terminal is a device such as a smartphone or PC that interfaces with the user through a chat application.

[1280] Generative AI

[1281] The generative AI receives the user's text through the server and recognizes the user's emotions using an emotion engine. The generative AI analyzes the text using natural language processing (NLP) techniques and generates an appropriate response. The generated response is sent to the terminal via the server and displayed to the user.

[1282] Examples

[1283] Example 1: Customer Service

[1284] A user uses a terminal to type, "Tell me about your recent order." This text is sent from the terminal to a server, which filters out certain words. The filtered text is sent to a generative AI, which uses an emotion engine to recognize the user's emotions. For example, if a user types text containing the word "delayed," the generative AI recognizes the emotion of anger and generates a response that reads, "We apologize for the inconvenience. We will review the current status of your order." This response is sent to the terminal via the server and displayed to the user.

[1285] Examples of prompt statements

[1286] "If a user enters text that expresses anger, how should a generative AI respond?"

[1287] By feeding this prompt into a generative AI model, it can learn how to generate an appropriate response that corresponds to the user's emotions.

[1288] Hardware and Software Used

[1289] Server: A powerful computer system (e.g. Linux server)

[1290] Devices: Smartphones, PCs

[1291] Generative AI: Generative AI models such as GPT-3

[1292] Protocol: HTTPS

[1293] Technology: Natural Language Processing (NLP), Emotion Recognition Engine

[1294] This system makes it possible to generate appropriate responses that correspond to the user's emotions while ensuring security.

[1295] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1296] Step 1:

[1297] A user uses the terminal to input text, for example, "Tell me about your recent orders." The input text is stored in the terminal's memory.

[1298] Step 2:

[1299] The terminal sends the text entered by the user to the server using the HTTPS protocol, and the data is encrypted with TLS. The input is the user's text, and the output is the encrypted text data.

[1300] Step 3:

[1301] The server filters the received text. If it contains certain words (e.g., customer information or confidential information), it blocks or masks that part. Regular expressions are used for filtering. The input is the encrypted text data, and the output is the filtered text data.

[1302] Step 4:

[1303] The server sends the filtered text to the generative AI, which includes an emotion engine to recognize the user's emotions. The input is the filtered text data, and the output is the emotion recognition result.

[1304] Step 5:

[1305] Generative AI recognizes the user's emotions from the received text. For example, if the text contains the word "delayed", it recognizes the emotion of anger. It uses natural language processing (NLP) techniques for emotion recognition. The input is filtered text data, and the output is the emotion recognition result.

[1306] Step 6:

[1307] Generative AI generates appropriate responses based on the recognized emotions. For example, if the user shows an angry emotion, it generates a calm and polite response. It uses a generative AI model such as GPT-3 to generate responses. The input is the emotion recognition result, and the output is the generated response.

[1308] Step 7:

[1309] The generative AI sends the generated response to the server. The server receives the response and sends it to the terminal. The input is the generated response, and the output is the encrypted response data.

[1310] Step 8:

[1311] The terminal displays the response it receives from the server to the user, for example, "We apologize for the inconvenience. We will check the current status of your order." The input is the encrypted response data, and the output is the text that is displayed to the user.

[1312] (Application example 2)

[1313] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[1314] Conventional customer service systems using generative AI had problems such as difficulty in ensuring security and responding appropriately to customer emotions. In addition, there was a lack of concrete measures to streamline customer service in brick-and-mortar stores. This could lead to a decrease in customer satisfaction and the risk of information leaks.

[1315] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1316] In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for recognizing the user's emotions using an emotion engine and adjusting the generative AI's response based on the emotions, and a means for supporting customer service in a physical store using a smart device. This makes it possible to provide appropriate responses according to the customer's emotions while ensuring security, and to improve the efficiency of customer service in a physical store.

[1317] The "Secure Proxy SaaS Service for Generative AI" is a cloud-based service that ensures security when using generative AI and provides governance such as blocking specific wording.

[1318] A "server" is a computer system that processes queries to the generative AI and performs security management and data filtering.

[1319] A "means of querying a generative AI via a server" is a method for sending data or questions to a generative AI via a server and receiving a response.

[1320] "Means of server management" are methods for maintaining the operation and security of a server, and for implementing governance such as blocking specific text.

[1321] "Measures to enable governance, such as blocking specific wording" refers to a filtering function to remove confidential information and inappropriate wording from the data processed by the generative AI.

[1322] An "emotion engine" is software that analyzes text input from a user and recognizes the emotions contained in that text.

[1323] A "smart device" is a portable electronic device with Internet connectivity, such as a smartphone or tablet.

[1324] "Means to support customer service in physical stores" refer to technologies and methods for making customer service in physical stores more efficient and improving customer satisfaction.

[1325] This invention is a system for improving customer satisfaction by streamlining customer service in brick-and-mortar stores. This system is composed of a secure proxy SaaS service for generative AI, a server, an emotion engine, and a smart device.

[1326] System Configuration

[1327] 1. Secure proxy SaaS service for generative AI:

[1328] The server provides a secure proxy SaaS service for generative AI. This service ensures security when using generative AI and provides governance such as blocking specific wording.

[1329] 2. Server:

[1330] The server will be installed within the company and will handle queries to the generative AI. The server will be managed using operational and security knowledge cultivated in the information and communications industry.

[1331] 3. Emotion Engine:

[1332] The emotion engine analyzes text input from users, recognizes the emotions contained within that text, and adjusts the generative AI's response based on the recognized emotions.

[1333] 4. Smart Devices:

[1334] Smart devices (such as smartphones and tablets) are used to assist in customer service in brick-and-mortar stores, allowing store staff to respond quickly and appropriately to customer inquiries.

[1335] System Operation

[1336] The server receives customer inquiries from smart devices and uses an emotion engine to recognize the emotion contained in the inquiry. The recognized emotion and the inquiry content are sent to the generative AI, which generates an appropriate response. The generated response is then sent back to the smart device after filtering to block certain phrases.

[1337] Hardware and Software Used

[1338] Hardware: Servers, smartphones, tablets

[1339] Software: Secure proxy SaaS service for generative AI, emotion recognition API, generative AI API

[1340] Examples

[1341] For example, if a customer asks, "Can you tell me more about this product?", the emotion engine will recognize the customer's interest and the generative AI will provide a detailed product explanation. If a customer asks, "I would like to return a product," the emotion engine will recognize the customer's dissatisfaction and the generative AI will provide a polite response.

[1342] Examples of prompt statements

[1343] "Please tell me more about this product"

[1344] "I'd like to return it."

[1345] "Do you have this item in stock?"

[1346] In this way, a system for supporting customer service in brick-and-mortar stores can be realized. This system provides appropriate responses according to the customer's emotions while ensuring security, making it possible to improve the efficiency of customer service in brick-and-mortar stores.

[1347] The flow of the specific process in the application example 2 will be described with reference to FIG.

[1348] Step 1:

[1349] A user enters a query using a smart device.

[1350] Input: User's query text (e.g. "Tell me more about this product")

[1351] Output: The query text is sent from the smart device to the server.

[1352] Step 2:

[1353] The server sends the received query text to an emotion engine to recognize emotions.

[1354] Input: Query text

[1355] Output: Perceived emotion (e.g. interest, anger, frustration)

[1356] Specific operation: The server calls the emotion recognition API and analyzes the query text to recognize emotions.

[1357] Step 3:

[1358] The server sends the recognized emotion and the query text to a generative AI to generate an appropriate response.

[1359] Input: query text, recognized sentiment

[1360] Output: The generated response text

[1361] How it works: The server calls the generative AI API, sends the query text and the recognized emotion as a prompt, and the generative AI generates a response.

[1362] Step 4:

[1363] The server filters the generated response text to block certain phrases.

[1364] Input: Generated response text

[1365] Output: The filtered response text

[1366] What happens: The server applies a filtering algorithm to filter out sensitive information and inappropriate language.

[1367] Step 5:

[1368] The server sends the filtered response text to the smart device.

[1369] Input: Filtered response text

[1370] Output: Response text displayed on smart device

[1371] Specific operation: The server sends the filtered response text to the smart device and displays it to the user.

[1372] Step 6:

[1373] The user checks the response on their smart device and makes additional enquiries if necessary.

[1374] Input: Filtered response text

[1375] Output: Additional query text from the user (e.g. "What other products do you recommend?")

[1376] Specific Actions: The user checks the response on the smart device and enters additional inquiries.

[1377] In this way, a system that supports customer service in brick-and-mortar stores is realized. The specific operations at each step ensure security and make it possible to provide appropriate responses according to the customer's emotions.

[1378] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input for the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1379] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with a prompt including an instruction, and also with voice data indicating a voice, a text message, and the like.

[1380] Inference data such as text data indicating a text and image data indicating an image are input. The data generation model 58 performs inference based on the input inference data in accordance with instructions indicated by a prompt, and outputs the inference result in a data format such as voice data and text data. Inference here refers to, for example, analysis, classification, prediction, and / or summarization.

[1381] Another example of generative AI is Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) are mentioned.

[1382] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1383] [Fourth embodiment]

[1384] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1385] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1386] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a wide area network (WAN) and / or a local area network (LAN).

[1387] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. In addition, the microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1388] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs the voice according to instructions from the processor 46.

[1389] Camera 42 is a small digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (e.g., an imaging range defined by an angle of view equivalent to the width of the field of vision of an average healthy person).

[1390] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54. The transmission and reception of various types of information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is performed in a secure state.

[1391] The control target 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, legs, etc. The posture and behavior of the robot 414 are controlled by controlling the motors of the arms, hands, legs, etc. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1392] Fig. 8 shows an example of main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1393] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32, and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1394] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1395] In the robot 414, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1396] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1397] "Example 1"

[1398] The present invention is a system that provides a secure proxy SaaS service for generative AI. Specifically, a server is installed within the company, and queries are made to the generative AI via this server. Server management utilizes operational and security knowledge cultivated in the information and communications industry, enabling governance such as blocking specific wording through settings for each individual company.

[1399] "Example 2"

[1400] As a concrete example, if a company wants to use generative AI to respond to customers, they can use the system of the present invention to make inquiries to the generative AI via a server installed within the company. In this case, server management can utilize the operational and security knowledge cultivated in information and communications to set up settings to block specific text, such as customer information or confidential information. This makes it possible to use generative AI in business while ensuring security.

[1401] The process flow of each embodiment will be described below.

[1402] "Example 1"

[1403] Step 1: Set up a server within the company. This server acts as an intermediary for communication with the generative AI.

[1404] Step 2: A query is sent to the generative AI via the server. This query is based on the company's business needs, for example, generating an automatic response for customer support.

[1405] Step 3: Manage the server. This management utilizes operational and security knowledge cultivated in the information and communications industry, enabling governance such as blocking specific wording with settings for each individual company.

[1406] "Example 2"

[1407] Step 1: If a company wants to use generative AI to handle customer support, it first sets up a server within the company.

[1408] Step 2: Make a query to the generative AI via the installed server. This query is

[1409] , generating automatic responses to customer inquiries, etc.

[1410] Step 3: Manage the server. This management utilizes operational and security knowledge cultivated in the information and communications industry, and sets up settings to block specific text, such as customer information or confidential information. This makes it possible to use generative AI in business while ensuring security.

[1411] Example 1

[1412] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1413] In order to use generative AI models safely and efficiently within a company, it is necessary to minimize the risk of data leakage to the outside while implementing governance based on the company's security policy. However, conventional systems lack the ability to block or filter specific text, making it difficult to fully comply with a company's security policy.

[1414] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1415] In this invention, the server includes a means for providing a secure proxy service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving a prompt text from a user's terminal, a means for inspecting the received prompt text based on filtering rules, a means for relaying the prompt text that has passed the inspection to the generative AI model, a means for receiving a response from the generative AI model and inspecting it again based on filtering rules, and a means for transmitting the response that has passed the filtering to the user's terminal. This makes it possible to safely and efficiently use the generative AI model while realizing governance based on the company's security policy.

[1416] The "Secure Proxy Service for Generative AI" is a service that securely relays queries to generative AI models and enables governance based on a company's security policies.

[1417] A “server” is a computer system installed within a company that relays queries to the generative AI model and provides filtering and governance functions.

[1418] A "prompt" is text that contains instructions or questions that a user enters into a generative AI model.

[1419] "Filtering rules" are rules or conditions for blocking specific words or phrases, and are set based on a company's security policy.

[1420] A "generative AI model" is an artificial intelligence model that generates responses based on prompts from the user.

[1421] "Governance" refers to management and control, such as blocking specific words or phrases, based on a company's security policy.

[1422] A "user's device" is a device, such as a computer or smartphone, that a user uses to input prompts to a generative AI model and receive responses.

[1423] An "inquiry" refers to instructions or questions that a user gives to a generative AI model, and is sent to the server as a prompt.

[1424] A "response" is an answer or piece of information that a generative AI model generates based on a prompt.

[1425] This invention is a system that provides a secure proxy service for generative AI, and queries generative AI models through a server installed within a company. A specific embodiment of this system will be described below.

[1426] Server installation and initial configuration

[1427] The server is installed within the company and connected to the company's network. The server administrator accesses the server's settings screen and sets filtering rules based on the company's security policy. For example, a rule is set to block inquiries that include the phrase "confidential information." Once the settings are complete, the server is connected to the company's network and begins operation.

[1428] User query input

[1429] The user inputs a prompt for the generative AI model from their own device. For example, a marketing person inputs a prompt such as, "Please tell me your sales forecast for the next quarter." Once the input is complete, the user clicks the send button.

[1430] The server receives and examines the query

[1431] The server receives the prompt text sent from the user's terminal. It checks the received prompt text based on the filtering rules. For example, it checks whether the text contains the phrase "confidential information." If there are no problems, it proceeds to the next step.

[1432] Relaying inquiries

[1433] The server relays the prompts that pass inspection to the generative AI model, which generates a response based on the prompts received. For example, it generates a response like, "Sales forecast for the next quarter is XXX yen."

[1434] Receiving and filtering responses by the server

[1435] The server receives the response from the generative AI model. The server checks the response again based on the filtering rules. For example, it checks whether the response contains the phrase "confidential information." If there are no problems, it proceeds to the next step.

[1436] Providing a response

[1437] The server sends the filtered responses to the user's device. The user can check the response from the generative AI model on the device. For example, the user can check the response, "Sales forecast for the next quarter is XXX yen."

[1438] Examples

[1439] As a concrete example, Company A has a setting to block queries that contain the phrase "confidential information." When a user enters the prompt phrase "Please generate a report that contains confidential information," the server detects this phrase and returns an error message without relaying the query.

[1440] Company B also uses a generative AI model to generate marketing reports. When a user enters a prompt such as "Please generate a report on the latest market trends," the server inspects the query and relays it to the generative AI model if it determines there is no problem. The response from the generative AI model is provided to the user via the server.

[1441] The system enables companies to use generative AI models safely and efficiently.

[1442] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1443] Step 1:

[1444] The server is installed within the company and connected to the network. The server administrator accesses the server's configuration screen and sets filtering rules based on the company's security policy. For example, a rule is set to block inquiries that include the phrase "confidential information." The input is the security policy and filtering rules, and the output is the configured server.

[1445] Step 2:

[1446] The user inputs a prompt sentence to the generative AI model from his / her own device. For example, the user inputs a prompt sentence such as "Please tell me your sales forecast for the next quarter." The input is the prompt sentence, and the output is the sent prompt sentence. The user clicks the send button.

[1447] Step 3:

[1448] The server receives the prompt text sent from the user's terminal. It checks the received prompt text based on filtering rules. For example, it checks whether it contains the phrase "confidential information." The input is the sent prompt text, and the output is the inspection result. If there are no problems, it proceeds to the next step.

[1449] Step 4:

[1450] The server relays the prompts that pass inspection to the generative AI model. The generative AI model generates a response based on the prompts it receives. For example, it generates a response like, "Sales forecast for the next quarter is XXX yen." The input is the prompt that passes inspection, and the output is the generated response.

[1451] Step 5:

[1452] The server receives the response from the generative AI model. The received response is inspected again based on the filtering rules. For example, it checks whether it contains the phrase "confidential information." The input is the response from the generative AI model, and the output is the inspection result. If there are no problems, it proceeds to the next step.

[1453] Step 6:

[1454] The server sends the filtered response to the user's device. The user can check the response from the generative AI model on the device. For example, the user can check the response "Sales forecast for the next quarter is XXX yen." The input is the filtered response, and the output is the response sent to the user's device.

[1455] (Application example 1)

[1456] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1457] When using generative AI within a company, there is a risk that confidential information may be leaked to the outside. In addition, it is difficult to implement appropriate governance when specific wording or data is included. Furthermore, security must be ensured when making inquiries to generative AI using smartphones.

[1458] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means. In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing a server, a means for making an inquiry to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for making an inquiry to the generative AI through an application installed on a smartphone, a means for securely communicating via a secure proxy server, a means for sending a filtered prompt to the generative AI, and a means for displaying the result on a smartphone application. This makes it possible to reduce the risk of confidential information leakage and perform appropriate governance when using generative AI within a company. In addition, high security can be ensured even when making inquiries to the generative AI using a smartphone.

[1459] The "Secure Proxy SaaS Service for Generative AI" is a cloud-based service that provides a proxy server for securely making inquiries to generative AI.

[1460] A "server" is a computer system that relays queries to the generative AI and manages security and governance.

[1461] A "means for querying a generative AI via a server" is a method for sending a prompt to a generative AI via a server and receiving a result.

[1462] "Means of server management" refers to the methods for maintaining and properly managing the operation and security of the server.

[1463] "Measures that enable governance, such as blocking specific wording" refers to methods for filtering out specific keywords and phrases and controlling the content of queries made to generative AI.

[1464] An "application installed on a smartphone" is software that runs on a smartphone and is used to make queries to the generative AI.

[1465] "Means for securely communicating via a secure proxy server" refers to a method for encrypting data and communicating securely through a secure proxy server.

[1466] "Means for transmitting a filtered prompt to the generative AI" refers to a method for transmitting the prompt after it has been filtered to the generative AI.

[1467] "Means for displaying results in a smartphone application" refers to a method for displaying the results returned from the generative AI in an application on a smartphone.

[1468] The following system configuration will be described as an embodiment of the present invention.

[1469] System Configuration

[1470] Hardware

[1471] Server: A computer system installed within a company that provides a secure proxy SaaS service for generative AI.

[1472] Smartphone: A device that allows users to make inquiries to generative AI.

[1473] software

[1474] Secure Proxy Server: Runs on the server, relays queries to the generative AI, and manages security and governance.

[1475] Smartphone application: Software that is installed on a smartphone and allows for queries to be made to the generative AI.

[1476] Data processing and calculation

[1477] server

[1478] The server provides a secure proxy SaaS service for generative AI. When a user makes a query to generative AI through a smartphone application, the prompt is first sent to the server. The server applies governance such as blocking certain words, and sends the filtered prompt to the generative AI.

[1479] Smartphone

[1480] A user uses a smartphone application to make a query to the generative AI. After entering a prompt and pressing the send button, the prompt is sent to the generative AI via a secure proxy server. The results returned from the generative AI are then displayed on the smartphone application via the secure proxy server again.

[1481] Examples

[1482] For example, when a company queries a generative AI for documents containing confidential information, the system can be used to reduce the risk of confidential information leaking to the outside. The user enters the following prompt into a smartphone application:

[1483] Prompt: "Generate the following sentence: Details about a company's new product."

[1484] This prompt is sent to a server where it is filtered to see if it contains specific words, and the filtered prompt is sent to a generative AI, which then displays the generated results on a smartphone application.

[1485] In this way, it is possible to reduce the risk of confidential information leaks and implement appropriate governance when using generative AI within a company. It is also possible to ensure high security when making inquiries to generative AI using smartphones.

[1486] The flow of the specific process in the application example 1 will be described with reference to FIG.

[1487] Step 1:

[1488] The user starts the smartphone application and inputs the question (prompt) they want to ask the generative AI. The input prompt is temporarily saved in the application.

[1489] Step 2:

[1490] When the user presses the send button, the smartphone application sends the prompt text to the secure proxy server. The input is the prompt text, and the output is a send request to the secure proxy server.

[1491] Step 3:

[1492] The server analyzes the received prompt text and filters it to see if it contains specific words. The input is the prompt text, and the output is the filtered prompt text. Specifically, the server checks it against a specific keyword list and removes any inappropriate words that it finds.

[1493] Step 4:

[1494] The server sends the filtered prompt to the generative AI. The input is the filtered prompt, and the output is a request to send to the generative AI. Specifically, the server sends an HTTP request to the API endpoint of the generative AI.

[1495] Step 5:

[1496] The generative AI generates a response based on the received prompt and sends the result back to the server. The input is the filtered prompt and the output is the generated response. In concrete terms, the generative AI uses its internal model to generate a response to the prompt.

[1497] Step 6:

[1498] The server receives the response returned from the generative AI and sends it to the smartphone application. The input is the generated response, and the output is a request to send to the smartphone application. In concrete terms, the server sends an HTTP request with the response data to the API endpoint of the smartphone application.

[1499] Step 7:

[1500] The smartphone application displays the response received from the server. The input is the generated response and the output is the result displayed to the user. In concrete terms, the application passes the received data to a UI component for display on the screen.

[1501] Example 2

[1502] Next, a second embodiment of the second embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1503] Customer support systems that use generative AI are required to respond to customers efficiently while ensuring security. However, with conventional systems, there is a risk that customer information or confidential information may be sent to the generative AI model, which may lead to security issues. In addition, responses from the generative AI model may be difficult for users to understand, which can lead to a decline in the quality of customer support. To solve these issues, a system that can filter the content of inquiries and convert the content of responses is needed.

[1504] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1505] In this invention, the server includes a means for receiving an inquiry from a terminal, a means for analyzing the received inquiry content and filtering out specific wording, a means for generating a prompt sentence based on the filtered inquiry content, a means for sending the generated prompt sentence to a generative AI model, a means for receiving a response from the generative AI model, a means for analyzing the received response and converting it into a format that is easy for a user to understand, a means for sending the converted response to the terminal, a means for installing a server, a means for making an inquiry to the generative AI model via the server, a means for managing the server, and a means for enabling governance such as blocking specific wording. This enables efficient customer service using generative AI while ensuring security.

[1506] A "terminal" is a device through which a user inputs inquiry content and transmits it to a server.

[1507] A "server" is a device that receives queries from a terminal, queries the generative AI model, and sends responses to the terminal.

[1508] "Inquiry content" is text data of a question or request that a user inputs via a terminal.

[1509] "Filtering" is the process of removing or masking specific wording (such as customer information or confidential information) from the content of an inquiry.

[1510] A "prompt sentence" is the text of a query that is generated to be sent to a generative AI model.

[1511] A "generative AI model" is an artificial intelligence model that generates a response based on an input prompt sentence.

[1512] A "response" is the text data of the answer returned from the generative AI model.

[1513] "Analysis" is the process of understanding the inquiry and response, and extracting or converting the necessary information.

[1514] "Transformation" is the process of changing the response from a generative AI model into a form that is easy for a user to understand.

[1515] "Governance" refers to the means of managing the operation and security of a system, and exercising control such as blocking specific text.

[1516] This invention relates to a customer service system that uses generative AI. Specifically, it is a system that uses a terminal, a server, and a generative AI model to filter inquiries and convert responses.

[1517] Hardware and Software Used

[1518] The terminal is a device where a user inputs the contents of an inquiry and transmits the input to the server. The terminal may be a PC, a smartphone, a tablet, or the like.

[1519] The server is a device that receives queries from the terminal, queries the generative AI model, and sends the response to the terminal. The server has installed software for parsing the query content, filtering, generating prompts, and parsing and converting the response. Specific software includes a natural language processing (NLP) library and an API client.

[1520] A generative AI model is an artificial intelligence model that generates responses based on input prompt sentences. Specifically, OpenAI's GPT-4 and other models are used.

[1521] Data processing and calculation

[1522] The terminal transmits the inquiry content input by the user to the server. For example, the user inputs "Please tell me the price of the new product."

[1523] The server analyzes the received inquiry content and filters out specific phrases (such as customer information or confidential information). Based on the filtered inquiry content, it generates a prompt sentence suitable for the generative AI model. For example, it generates a prompt sentence such as "Please tell me the price of the new product."

[1524] The server sends the generated prompt sentence to the generative AI model, which generates a response based on the prompt sentence and sends it back to the server. For example, it generates a response such as "The price of the new product is $100."

[1525] The server parses the response and converts it into a format that is easy for the user to understand. It then sends the converted response to the terminal. The terminal displays the response to the user. For example, the message "The price of the new product is $100" is displayed on the user's screen.

[1526] Examples

[1527] For example, if a company receives an inquiry from a customer saying, "Please tell me the price of a new product," the following process will take place:

[1528] 1. On the terminal, the user enters "Please tell me the price of the new product" and clicks the send button.

[1529] 2. The server analyzes the received data and extracts the text "Please tell me the price of the new product." It then performs a filtering process to check whether the data contains a specific phrase.

[1530] 3. The server generates a prompt sentence based on the filtered text: "Please tell me the price of the new product."

[1531] 4. The server creates and sends an API request to send the generated prompt to GPT-4.

[1532] 5. The server receives the response from GPT-4 and gets the text “The price of the new product is $100.”

[1533] 6. The server parses the received text and replaces technical terms with more common language where necessary.

[1534] 7. The server sends the converted response to the terminal, which displays the received response to the user, and the message "The new product costs $100" appears on the user's screen.

[1535] In this way, businesses can leverage generative AI to deliver efficient customer service while ensuring security.

[1536] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1537] Step 1:

[1538] The terminal transmits the inquiry content entered by the user to the server. As input, the user inputs "Please tell me the price of the new product" and clicks the send button. As output, the inquiry content is transmitted to the server.

[1539] Step 2:

[1540] The server analyzes the inquiry received from the terminal. As input, it receives the text data "Please tell me the price of the new product" sent from the terminal. To process the data, it analyzes the text using natural language processing (NLP) technology and filters out specific phrases (such as customer information or confidential information). As output, it obtains the filtered inquiry content.

[1541] Step 3:

[1542] The server generates a prompt sentence based on the filtered inquiry content. As input, it receives the filtered text "Please tell me the price of the new product." As data processing, it converts it into a format suitable for the generative AI model. As output, it obtains the generated prompt sentence.

[1543] Step 4:

[1544] The server sends the generated prompt sentence to the generative AI model. As input, it receives the generated prompt sentence "Please tell me the price of the new product." As a specific operation, it creates an API request and sends it to the generative AI model (e.g., GPT-4). As output, the prompt sentence is sent to the generative AI model.

[1545] Step 5:

[1546] The server receives the response from the generative AI model. As input, it receives the response data returned from the generative AI model. As a specific operation, it parses the API response and obtains the text "The price of the new product is $100." As output, it obtains the response from the generative AI model.

[1547] Step 6:

[1548] The server analyzes the received response and converts it into a format that is easy for the user to understand. As input, it receives the response data "The price of the new product is $100." As data processing, it performs conversion processing such as replacing technical terms with common terms. As output, it obtains the converted response.

[1549] Step 7:

[1550] The server sends the converted response to the terminal. As input, it receives the converted response data "The price of the new product is $100." As a specific operation, it sends the response data to the terminal. As output, the response is sent to the terminal.

[1551] Step 8:

[1552] The terminal displays the received response to the user. As input, it receives the response data "The price of the new product is $100" sent from the server. As a specific operation, it displays the response message on the user's screen. As output, the user can confirm the response.

[1553] (Application example 2)

[1554] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1555] Conventional customer service systems using generative AI have difficulty in ensuring security and cannot block certain text (e.g., customer information or confidential information), making it difficult to use safely within a company. In addition, they lacked the functionality to make personalized product recommendations to users, making it difficult to improve customer satisfaction.

[1556] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1557] In this invention, the server includes a means for providing a secure proxy service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving questions from users and sending them to the generative AI, a means for returning answers from the generative AI to the user, and a means for making personalized product recommendations based on the user's purchase history and browsing history. This enables safe customer service and personalized product recommendations using generative AI while ensuring security.

[1558] The "Secure Proxy Service for Generative AI" is a relay service that enables safe inquiries to be made to generative AI, and exchanges data while ensuring security.

[1559] A "server" is a computer system that stores, processes, and manages data on a network.

[1560] A "means for querying a generative AI via a server" is a method for sending questions or instructions to a generative AI via a server.

[1561] "Server management" refers to the activity of operating, maintaining, and managing the security of a server.

[1562] "Governance such as blocking specific wording" is a management technique that filters out specific keywords and phrases (such as customer information or confidential information) to prevent the leakage of inappropriate information within the system.

[1563] "Means for receiving questions from a user and sending them to a generative AI" refers to a method for receiving questions entered by a user and sending them to a generative AI.

[1564] "Means of returning an answer from a generative AI to a user" refers to a method for returning an answer obtained from a generative AI to a user.

[1565] "Means for personalized product recommendations based on a user's purchase history and browsing history" refers to a method for analyzing a user's past purchase history and browsing history and recommending individually optimized products based on that.

[1566] A system for implementing this invention includes a server that provides a secure proxy service for the generative AI, a terminal that receives questions from users and sends them to the generative AI, a terminal that returns answers from the generative AI to the user, and a function that makes personalized product recommendations based on the user's purchasing history and browsing history.

[1567] Hardware and Software Configuration

[1568] Server: Installed within a company, it provides a secure proxy service for generative AI. The server is built using a web framework such as Flask.

[1569] Device: A device used by a user, such as a smartphone or tablet. These devices run applications that receive questions from users and send them to the generative AI.

[1570] Generative AI: Use generative AI models such as OpenAI's GPT-3. Generative AI generates appropriate answers to questions posed by users.

[1571] Data processing and calculation

[1572] 1. Receiving user questions:

[1573] The user inputs a question through the terminal, for example, "How do you use this product?"

[1574] 2. Filter questions:

[1575] The server filters certain phrases (customer information and confidential information) from the question, ensuring security.

[1576] 3. Querying the generative AI:

[1577] The filtered questions are sent to a generative AI, which generates answers to the questions.

[1578] 4. Response to the answer:

[1579] The server receives the answer from the generative AI and responds to the user through the device, for example, "This product is used in the following steps..."

[1580] 5. Personalized product recommendations:

[1581] The server analyzes the user's purchase and browsing history and recommends individually optimized products based on that information, for example by displaying "These are the products we recommend for you."

[1582] Examples

[1583] When a user asks, "How do you use this product?", the server filters the question and sends it to the generative AI. The generative AI responds with, "This product is used in the following steps..." and the answer is returned to the user. Related products are also recommended based on the user's past purchase and browsing history.

[1584] Examples of prompt statements

[1585] "How do you use this product?"

[1586] In this way, secure customer interactions and personalized product recommendations using generative AI can be achieved while ensuring security.

[1587] The flow of the specific process in the application example 2 will be described with reference to FIG.

[1588] Step 1:

[1589] A user inputs a question through a terminal, for example, "How do you use this product?" The input question is sent to a server by an application on the terminal.

[1590] Step 2:

[1591] The server filters the questions it receives. Specifically, it detects specific phrases (customer information or confidential information) from the questions and blocks or masks them. The filtered questions are then sent to the generative AI.

[1592] Step 3:

[1593] The server sends the filtered questions to the generative AI, which generates appropriate answers to the questions it receives. For example, it uses OpenAI's GPT-3 to generate an answer like, "This product is used in the following steps..."

[1594] Step 4:

[1595] The server receives the answer from the generative AI. The server returns the answer to the device. The device displays the answer from the generative AI to the user.

[1596] Step 5:

[1597] The server analyzes the user's purchase history and browsing history. Specifically, it retrieves the user's past purchase history and browsing history from the database and makes personalized product recommendations based on that. For example, it displays "These are the products we recommend for you."

[1598] Step 6:

[1599] The server sends the personalized product recommendations to the terminal. The terminal displays the recommended products to the user. The user can review the recommended products and consider purchasing them.

[1600] In this way, secure customer interactions and personalized product recommendations using generative AI can be achieved while ensuring security.

[1601] In addition, an emotion engine that estimates the emotion of the user may be further combined. That is, the identification processing unit 290 may estimate the emotion of the user using the emotion identification model 59, and perform identification processing using the emotion of the user.

[1602] "Example 1"

[1603] As one embodiment of the present invention, there is a system that provides a secure proxy SaaS service for generative AI, which is combined with an emotion engine. This system queries the generative AI through a server installed in the company and manages the server. The server management utilizes operational and security knowledge cultivated in information and communications, enabling governance such as blocking specific wording. Furthermore, the generative AI includes an emotion engine that recognizes the user's emotions. Specifically, it receives text input from the user and recognizes the user's emotions from the text. Then, the generative AI's response is adjusted based on the recognized emotion. For example, if a user inputs text indicating an emotion of anger, the emotion engine recognizes the anger, and the generative AI generates an appropriate response based on the recognition result. This enables a more appropriate response that corresponds to the user's emotions.

[1604] "Example 2"

[1605] As one embodiment of the present invention, there is a system that provides a secure proxy SaaS service for generative AI, which is combined with an emotion engine. This system queries the generative AI through a server installed in the company and manages the server. The server management utilizes operational and security knowledge cultivated in information and communications, enabling governance such as blocking specific wording. Furthermore, the generative AI includes an emotion engine that recognizes the user's emotions. Specifically, it receives text input from the user and recognizes the user's emotions from the text. Then, the generative AI's response is adjusted based on the recognized emotion. For example, if a user inputs text indicating an emotion of anger, the emotion engine recognizes the anger, and the generative AI generates an appropriate response based on the recognition result. This enables a more appropriate response that corresponds to the user's emotions.

[1606] The process flow of each embodiment will be described below.

[1607] "Example 1"

[1608] Step 1: Receive text input from a user, which may include the user's sentiment.

[1609] Step 2: Input the received text into the emotion engine, which recognizes the user's emotion from the text.

[1610] Step 3: The output from the emotion engine (the recognized emotion) is fed into the generative AI, which generates a response based on this emotion.

[1611] Step 4: The generative AI's response is returned to the user. This response corresponds to the user's emotion.

[1612] Example 1

[1613] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1614] Conventional services using generative AI have security and governance issues, making it difficult to block specific text or operate based on a company's security policy. In addition, they are unable to generate responses that take into account the user's emotions, which can lead to a poor user experience.

[1615] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1616] In this invention, the server includes a means for providing a secure proxy service for the generative AI, a means for installing the server, a means for making an inquiry to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific words, a means for including an emotion engine for recognizing the user's emotions, and a means for adjusting the response of the generative AI based on the emotions recognized by the emotion engine. This makes it possible to generate an appropriate response corresponding to the user's emotions while strengthening security and governance.

[1617] "Generative artificial intelligence" is an artificial intelligence technology that generates natural language responses and content based on user input.

[1618] A "secure proxy service" is a service that relays user requests and provides proxy functions to enhance security and governance.

[1619] A "server" is a computer system that processes and manages data on a network.

[1620] "Governance" refers to the management and control of system operations, such as blocking specific wording and applying security policies.

[1621] An "emotion engine" is software or hardware for analyzing text input from a user and recognizing its emotions.

[1622] "Adjusting the response" means changing the content of the response generated by the generative artificial intelligence based on the emotion recognized by the emotion engine.

[1623] The present invention is a system that provides a secure proxy service for querying a generative artificial intelligence through a server installed within a company. The system includes an emotion engine that recognizes the user's emotions and adjusts responses based on the emotions.

[1624] Hardware and Software Used

[1625] 1. Server

[1626] It is installed within a company.

[1627] It will be managed utilizing operational and security knowledge cultivated in the information and communications industry.

[1628] It has a governance function that blocks specific wording.

[1629] 2. Terminal

[1630] A device (e.g. PC, smartphone) that allows a user to input text.

[1631] 3. Generative Artificial Intelligence

[1632] An artificial intelligence technology that generates natural language responses based on user input.

[1633] 4. Emotion Engine

[1634] Software that analyzes text input from users and recognizes their emotions.

[1635] Data processing and calculation

[1636] 1. User text input

[1637] A user uses the terminal to input text.

[1638] For example: "Why is it so slow?"

[1639] 2. Receiving and parsing the text by the server

[1640] The server receives text input from the user and checks whether it contains specific words.

[1641] Example: Detect the keyword "slow".

[1642] 3. Emotion Recognition by Emotion Engine

[1643] The server sends the text to an emotion engine, which analyzes the text to recognize the user's emotion.

[1644] Example: The emotion engine recognizes "anger."

[1645] 4. Response generation using generative artificial intelligence

[1646] The server sends the emotion recognition results to the generative artificial intelligence, which then generates an appropriate response.

[1647] For example: "We apologize for the inconvenience. We are currently working on improving our system."

[1648] 5. Server Sends Response

[1649] The server sends the generated response to the user's terminal.

[1650] Examples

[1651] For example, if a user enters the text "Why is it so slow?", the following happens:

[1652] 1. User text input

[1653] A user types in the chat window, "Why is this so slow?"

[1654] 2. Receiving and parsing the text by the server

[1655] The server receives the text "Why is it so slow?" and detects the keyword "slow".

[1656] 3. Emotion Recognition by Emotion Engine

[1657] The server sends the text to an emotion engine, which recognizes "anger."

[1658] 4. Response generation using generative artificial intelligence

[1659] The server sends the emotion recognition results to the generative artificial intelligence, which generates a response saying, "We apologize for the inconvenience. We are currently working on improving the system."

[1660] 5. Server Sends Response

[1661] The server displays the generated response in the user's chat window.

[1662] Examples of prompt statements

[1663] Here are some example prompts to input to a generative AI model:

[1664] User: Why is it so slow?

[1665] Emotion engine: Anger recognition

[1666] Generative AI: We apologize for the inconvenience. We are currently working on improving the system.

[1667] In this way, a more appropriate response can be made according to the user's emotions.

[1668] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1669] Step 1:

[1670] Initial Server Setup

[1671] The server is installed within the company. The server administrator accesses the server's settings screen and sets filtering rules and security policies. Specifically, they set rules to block specific wording and apply security policies. The input is the administrator's settings information, and the output is the state of the server with the settings reflected.

[1672] Step 2:

[1673] User text input

[1674] A user uses a terminal to input text, specifically, typing "why is this so slow?" into a chat window. The input is the user's text, and the output is the transmission of text data from the terminal to the server.

[1675] Step 3:

[1676] Receiving and parsing text by the server

[1677] The server receives text input from the user. It analyzes the received text and checks whether it contains a specific phrase. Specifically, it detects the keyword "slow." The input is the user's text data, and the output is the analysis result.

[1678] Step 4:

[1679] Emotion recognition by emotion engine

[1680] The server sends the text to the emotion engine. The emotion engine analyzes the text and recognizes the user's emotion. Specifically, the emotion engine recognizes "anger." The input is the analyzed text data, and the output is the emotion recognition result.

[1681] Step 5:

[1682] Response generation using generative artificial intelligence

[1683] The server sends the emotion recognition results to the generative AI. The generative AI generates an appropriate response based on the emotion recognition results. Specifically, it generates a response such as "We apologize for the inconvenience. We are currently working on improving the system." The input is the emotion recognition results, and the output is the generated response.

[1684] Step 6:

[1685] Server sends response

[1686] The server sends the generated response to the user's terminal, and displays the generated response in a chat window. The input is the generated response data, and the output is the response displayed on the user's terminal.

[1687] (Application example 1)

[1688] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1689] Conventional generative AI systems were unable to respond in a way that took the user's emotions into account, and had insufficient functions for blocking specific phrases and for governance. This resulted in security risks and a poor user experience when used within a company. It was also difficult to provide appropriate responses when used on mobile devices such as smartphones.

[1690] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1691] In this invention, the server includes a means for providing a secure proxy SaaS service for generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific words, a means for recognizing the user's emotions using an emotion engine and adjusting the generative AI's response based on the emotion, and a means for functioning as an application installed on a smartphone. This makes it possible to reduce security risks within a company and provide appropriate responses according to the user's emotions.

[1692] The "Secure Proxy SaaS Service for Generative AI" is a Software as a Service (SaaS) that provides a proxy service for securely making inquiries to generative artificial intelligence (AI).

[1693] "Server" refers to a computer system that runs the secure proxy SaaS service for generative AI and processes and manages data within the company.

[1694] A "means of querying a generative AI via a server" is a method for sending data or requests to a generative AI via a server and receiving a response.

[1695] "Means of server management" refers to methods for managing the operation and security of servers and providing stable services.

[1696] "Measures to enable governance, such as blocking specific wording" refers to a method that provides the ability to detect and block specific keywords or phrases based on a company's security policy.

[1697] An "emotion engine" is a software component for analyzing emotions from a user's input text and generating an appropriate response based on the emotions.

[1698] "Means to function as an application installed on a smartphone" refers to software that operates on a smartphone and is provided in a form that the user can use.

[1699] The system for implementing this invention is composed mainly of a server that provides a secure proxy SaaS service for generative AI. The server is installed within a company and provides a proxy service for securely making inquiries to generative AI. The server utilizes operational and security knowledge cultivated in information and communications to enable governance such as blocking specific wording.

[1700] The server is equipped with an emotion engine that recognizes the user's emotions, receives text input from the user, analyzes the emotions from the text, and uses the TextBlob library to obtain the polarity of the emotions and adjusts the generative AI's response based on the user's emotions.

[1701] Users access the system through an application installed on their smartphones, which sends user input to the server and receives responses from it, which checks for specific phrases, generates prompts based on sentiment analysis, and uses the OpenAI API to generate appropriate responses.

[1702] For example, if a user enters "I'm very tired today," the emotion engine recognizes that the user is tired (a negative emotion). The server generates a prompt sentence like this:

[1703] "The user is tired. Generate an appropriate response to the following message: I'm very tired today."

[1704] Based on this prompt, a response is generated using the OpenAI API, and provided to the user, for example, "Thank you for your hard work. Please take it easy and get some rest today."

[1705] This system can reduce security risks within a company and provide appropriate responses according to the user's emotions.

[1706] The flow of the specific process in the application example 1 will be described with reference to FIG.

[1707] Step 1:

[1708] A user starts an application installed on a smartphone and inputs text. The input text is a sentence in a natural language that includes the user's emotions and intentions. The input data is sent from the user's smartphone to a server.

[1709] Step 2:

[1710] The server parses the received text input and checks for specific terms. It checks whether the text contains any of the keywords or phrases in the list of specific terms. If the text contains any of the specific terms, the server generates a warning message and sends it back to the user. The input data is the text input, and the output data is either a warning message or text input for the next step.

[1711] Step 3:

[1712] If no specific phrase is included, the server performs sentiment analysis on the text input using the sentiment engine. It uses the TextBlob library to calculate the sentiment polarity (positive, negative, neutral) of the text. The input data is the text input, and the output data is the sentiment polarity.

[1713] Step 4:

[1714] The server generates a prompt sentence based on the result of the sentiment analysis. For example, if the sentiment polarity is negative, the server creates a prompt sentence in the form of "The user is tired. Please generate an appropriate response to the following message: [Text input]". The input data is the sentiment polarity and the text input, and the output data is the prompt sentence.

[1715] Step 5:

[1716] The server sends the generated prompt to the OpenAI API, which generates an appropriate response. The OpenAI API generates a natural language response based on the prompt and sends it back to the server. The input data is the prompt, and the output data is the generated response.

[1717] Step 6:

[1718] The server sends the generated response to the user's smartphone. The user receives and displays the response through an application on the smartphone. The input data is the generated response, and the output data is the response displayed on the user's smartphone.

[1719] Through the above processing steps, it is possible to provide an appropriate response according to the user's emotions.

[1720] Example 2

[1721] Next, a second embodiment of the second embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1722] Conventional systems using generative AI were unable to respond in a way that took the user's emotions into account, and there were also security concerns. In particular, when used within a company, there was a high risk of customer information and confidential information being leaked, and appropriate governance was required to prevent this. In addition, because it was not possible to generate appropriate responses based on the user's emotions, it was difficult to improve customer satisfaction.

[1723] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1724] In this invention, the server includes a means for providing a secure proxy service for the generative AI, a means for installing the server, a means for making inquiries to the generative AI via the server, a means for managing the server, a means for enabling governance such as blocking specific wording, a means for receiving text input from a user and sending the text to the generative AI, a means for the generative AI to recognize the user's emotions and generate a response based on the emotions, and a means for returning the generated response to the user. This makes it possible to generate an appropriate response corresponding to the user's emotions while ensuring security.

[1725] The "Secure Proxy Service for Generative AI" is a relay service that allows secure access to generative AI and has the ability to filter and encrypt data.

[1726] A "server" is a computer system that provides services to other computers over a network.

[1727] "Means of querying the generative AI via a server" refers to a function for sending data to the generative AI via a server and receiving the response.

[1728] "Server management" refers to the set of activities for operating, monitoring, and maintaining a server, including security measures and performance optimization.

[1729] "Governance such as blocking specific wording" is a management technique for preventing information leaks by detecting specific keywords or phrases and blocking or masking them.

[1730] "Means for receiving text input from a user and sending that text to the generative AI" refers to a function for receiving text data entered by a user and sending it to the generative AI.

[1731] "Means for the generative AI to recognize the user's emotions and generate a response based on those emotions" refers to the function that enables the generative AI to analyze emotions from the user's input text and generate an appropriate response based on the results of that analysis.

[1732] "Means of returning the generated response to the user" refers to a function for sending the response generated by the generative AI to the user via a server and displaying it to the user.

[1733] The present invention relates to a system for providing a secure proxy service for generative AI. A specific embodiment of this system will be described below.

[1734] System Configuration

[1735] server

[1736] The server is installed within the company and manages inquiries to the generative AI. The server utilizes operational and security knowledge cultivated in the information and communications industry to set up the system to block specific text (e.g., customer information or confidential information). The server uses the HTTPS protocol to encrypt data and ensure security.

[1737] Terminal

[1738] The terminal receives text input from the user and transmits the text to the server. The terminal is a device such as a smartphone or PC that interfaces with the user through a chat application.

[1739] Generative AI

[1740] The generative AI receives the user's text through the server and recognizes the user's emotions using an emotion engine. The generative AI analyzes the text using natural language processing (NLP) techniques and generates an appropriate response. The generated response is sent to the terminal via the server and displayed to the user.

[1741] Examples

[1742] Example 1: Customer Service

[1743] A user uses a terminal to type, "Tell me about your recent order." This text is sent from the terminal to a server, which filters out certain words. The filtered text is sent to a generative AI, which uses an emotion engine to recognize the user's emotions. For example, if a user types text containing the word "delayed," the generative AI recognizes the emotion of anger and generates a response that reads, "We apologize for the inconvenience. We will review the current status of your order." This response is sent to the terminal via the server and displayed to the user.

[1744] Examples of prompt statements

[1745] "If a user enters text that expresses anger, how should a generative AI respond?"

[1746] By feeding this prompt into a generative AI model, it can learn how to generate an appropriate response that corresponds to the user's emotions.

[1747] Hardware and Software Used

[1748] Server: A powerful computer system (e.g. Linux server)

[1749] Devices: Smartphones, PCs

[1750] Generative AI: Generative AI models such as GPT-3

[1751] Protocol: HTTPS

[1752] Technology: Natural Language Processing (NLP), Emotion Recognition Engine

[1753] This system makes it possible to generate appropriate responses that correspond to the user's emotions while ensuring security.

[1754] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1755] Step 1:

[1756] A user uses the terminal to input text, for example, "Tell me about your recent orders." The input text is stored in the terminal's memory.

[1757] Step 2:

[1758] The terminal sends the text entered by the user to the server using the HTTPS protocol, and the data is encrypted with TLS. The input is the user's text, and the output is the encrypted text data.

[1759] Step 3:

[1760] The server filters the received text. If it contains certain words (e.g., customer information or confidential information), it blocks or masks that part. Regular expressions are used for filtering. The input is the encrypted text data, and the output is the filtered text data.

[1761] Step 4:

[1762] The server sends the filtered text to the generative AI, which includes an emotion engine to recognize the user's emotions. The input is the filtered text data, and the output is the emotion recognition result.

[1763] Step 5:

[1764] Generative AI recognizes the user's emotions from the received text. For example, if the text contains the word "delayed", it recognizes the emotion of anger. It uses natural language processing (NLP) techniques for emotion recognition. The input is filtered text data, and the output is the emotion recognition result.

[1765] Step 6:

[1766] Generative AI generates appropriate responses based on the recognized emotions. For example, if the user shows an angry emotion, it generates a calm and poli...

Claims

1. A means for receiving inquiry content from a terminal; A means for analyzing the received inquiry content and filtering out specific phrases; means for generating a prompt sentence based on the filtered query content; A means for transmitting the generated prompt sentence to a generative AI model; means for receiving a response from the generative AI model; means for parsing said received response and converting it into a user understandable format; means for transmitting the converted response to the terminal; A system including:

2. means for recognizing emotions of the user using an emotion engine and adjusting a response of a generative AI model based on the recognized emotions; The system of claim 1 .

3. A means for providing personalized product recommendations based on at least one of the user's purchase history and browsing history, The system of claim 1 .

Images