System

Abstract

A system is provided.SOLUTION: A system comprising: user input means; natural language processing means for analyzing a query from the user input means; database connection means for retrieving related information based on the query analyzed by the natural language processing means; natural language generation means for generating a user-friendly answer based on the information acquired by the database connection means; and display means for presenting the answer generated by the natural language generation means to the user.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 an instruction sentence related to a description of 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] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] As generative AI becomes more widespread, the risk of copyright infringement and security breaches increases even when it is used within companies. However, many companies have not kept up with legal regulations and do not have an environment in which they can easily check appropriate information regarding the use of generative AI, making the psychological hurdles to utilizing generative AI high. To solve this issue, there is a need for a system that provides an environment in which companies can use generative AI with peace of mind and improves work efficiency. [Means for solving the problem]

[0005] The present invention solves the above-mentioned problems by providing a system including a user input means, a natural language processing means for analyzing a query from the user input means, a database connection means for searching for related information based on the query analyzed by the natural language processing means, a natural language generation means for generating an answer that is easy for the user to understand based on the information acquired by the database connection means, and a display means for presenting the answer generated by the natural language generation means to the user. This allows users to quickly obtain appropriate answers when they have questions about the introduction or use of generative AI, and enables companies to quickly obtain guidelines for using generative AI with confidence.

[0006] The "user input means" is an interface through which a user inputs a query and sends it to the system.

[0007] "Natural language processing means" refers to a method and module that analyzes queries entered by users and extracts keywords and contextual information.

[0008] The "database connection means" is a module for searching and obtaining related information from a database based on the analyzed query.

[0009] "Natural language generation means" refers to a method and module for generating answers that are easy for users to understand based on acquired information.

[0010] The "display means" is an interface for displaying the generated answers so that the user can check them. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 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. [Figure 3]FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 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. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 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. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0012] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0013] First, the terms used in the following description will be explained.

[0014] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single 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), and an APU (Accelerated Processing Unit).

[0015] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0016] 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.

[0017] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0018] 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. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0019] [First embodiment]

[0020] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0021] 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.

[0022] 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 WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0023] 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.

[0024] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. 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.

[0025] 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 (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0026] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0027] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0028] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. 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 process 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.

[0029] 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.

[0030] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. 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.

[0031] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. 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."

[0032] The security concierge system of the present invention is a system for reducing security and legal risks associated with the use of generative AI in companies, allowing companies to use generative AI with peace of mind. This system includes, as its main components, a user input means, a natural language processing means, a database connection means, a natural language generation means, and a display means.

[0033] System configuration

[0034] 1. User Input Method

[0035] Users access the web interface via a terminal.

[0036] This interface includes a question form and a submit button, allowing users to enter and submit questions about the generated AI.

[0037] 2. Natural Language Processing Methods

[0038] The server receives a question sent from a user input means.

[0039] The received question is analyzed by a natural language processing module to extract the most important keywords and context.

[0040] 3. Database connection method

[0041] The server uses the keywords extracted by natural language processing to send queries to a database to search for relevant information.

[0042] This database stores information about corporate security rules and regulations.

[0043] 4. Natural language generation means

[0044] The server generates an answer that is easy for the user to understand based on the information retrieved from the database.

[0045] A natural language generation module takes over this process to generate the most appropriate answer to the query.

[0046] 5. Display means

[0047] The server sends the generated answer to the user's terminal.

[0048] The terminal displays the received answers on a web interface for easy confirmation by the user.

[0049] Specific examples

[0050] 1. The user accesses the Security Concierge web interface on their device.

[0051] 2. The user enters the question "I would like to use ChatGPT (registered trademark) in my company, but are there any security issues?" into the question form and presses the submit button.

[0052] 3. The server receives the question, and the natural language processing module extracts the key keywords "ChatGPT" and "security."

[0053] 4. Based on these keywords, the server searches the database for relevant company rules and legal information.

[0054] 5. The database returns the search results (e.g., the company's security policy regarding the use of ChatGPT).

[0055] 6. The server uses a natural language generation module to generate a response that states, "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, and 2. Regular security reviews are conducted."

[0056] 7. The server formats this response in HTML or other format and sends it to the terminal.

[0057] 8. The terminal displays the received answer to the user, who can then take necessary action based on the answer.

[0058] In this way, the system can quickly and accurately answer user questions about the use of generative AI and provide guidelines for companies to use generative AI with confidence.

[0059] The processing flow will be explained below.

[0060] Step 1:

[0061] The user opens a web browser on the device and accesses the Security Concierge web interface.

[0062] Step 2:

[0063] Users enter questions about the generative AI service into a question form on the interface (e.g., "I would like to use ChatGPT in my company, but are there any security issues?").

[0064] Step 3:

[0065] After the user enters the question, he or she clicks the submit button.

[0066] Step 4:

[0067] The terminal sends the user's input to the server as an HTTP request.

[0068] Step 5:

[0069] The server receives the HTTP request and extracts the user's question.

[0070] Step 6:

[0071] The server passes the question to a natural language processing (NLP) module to begin parsing it.

[0072] Step 7:

[0073] The natural language processing (NLP) module analyzes the question and extracts keywords and contextual information (e.g., "ChatGPT," "security").

[0074] Step 8:

[0075] The server generates a search query based on the extracted keywords.

[0076] Step 9:

[0077] The server sends the generated search query to the database.

[0078] Step 10:

[0079] The database searches for relevant company rules and legal information based on the received query.

[0080] Step 11:

[0081] The database returns the search results to the server.

[0082] Step 12:

[0083] The server passes the retrieved information to a natural language generation (NLG) module to begin generating an answer.

[0084] Step 13:

[0085] The Natural Language Generation (NLG) module generates answers based on the search results in a user-friendly format (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted").

[0086] Step 14:

[0087] The server converts the generated response into HTML format and sends it to the user's terminal.

[0088] Step 15:

[0089] The terminal displays the received HTML data in a web browser.

[0090] Step 16:

[0091] The user confirms the answers on the screen and decides on the next action (e.g., setting restrictions on access outside the company network).

[0092] Example 1

[0093] Next, a description will be given of 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."

[0094] The use of generative AI in companies involves security and legal risks, making it difficult to use safely. This can delay the introduction of generative AI or cause problems after its introduction. The objective of this invention is to provide a system for reducing these risks and operating generative AI safely and efficiently.

[0095] 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.

[0096] In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, a means for the user input means to receive a user input via a web interface and send a query by operating a send button, a means for the natural language processing means to analyze keywords and context information in the query using a natural language processing module, a means for the natural language generation means to generate an answer based on information obtained from a database using the natural language generation module, and a means for the display means to receive the generated answer and display it on the web interface. This reduces security and legal risks when using generative AI, allowing companies to use generative AI with peace of mind.

[0097] "User input means" refers to a means by which a user provides input to the system.

[0098] "Natural language processing means" is a means for analyzing queries received from users and extracting important keywords and contextual information.

[0099] The "database connection means" is a means for accessing a database based on a query and searching for related information.

[0100] "Natural language generation means" is a means for generating appropriate answers to users based on acquired information.

[0101] The "display means" is a means for presenting the generated answer to the user.

[0102] A "web interface" is an interface through which a user interacts with a system via a web browser.

[0103] A "natural language processing module" is a software module for analyzing input text and extracting keywords and contextual information.

[0104] A "natural language generation module" is a software module for generating answers in natural language based on analyzed information.

[0105] A "server" is a computer that performs the central processing of the system, and is a device that receives input from users and analyzes, processes, and generates responses.

[0106] A "query" refers to a question or request made by a user to a system.

[0107] The security concierge system of the present invention is intended to enable the safe use of generative AI in companies and to reduce security and legal risks. This system is implemented using the following hardware and software.

[0108] User Input Method

[0109] A user accesses the web interface of the security concierge system using a web browser. This interface includes a question form and a submit button that serves to transmit the query entered by the user to the server.

[0110] Natural language processing tools

[0111] The server receives queries sent by users. This processing is performed using web server software such as Nginx or Apache (registered trademark). The received queries are processed using web frameworks implemented in Python, such as Django or Flask. The server then calls a natural language processing module to analyze the queries. This module uses natural language processing libraries such as NLTK or spaCy.

[0112] Database connection method

[0113] The server uses the keywords and contextual information extracted from the query to send an SQL query to a database. This database stores information on corporate security rules and regulations, and uses an RDBMS such as MySQL (registered trademark) or PostgreSQL. The server accesses the database using a database connection library (e.g., SQLAlchemy or Psycopg2) to retrieve the relevant information.

[0114] Natural language generation means

[0115] The server uses a natural language generation module based on the information retrieved from the database to generate answers for the user. This module uses, for example, OpenAI's GPT-3 (registered trademark) or GPT-4 (registered trademark). Answers are generated in a format that is easy for the user to understand.

[0116] Display means

[0117] The server formats the generated response in HTML format or similar and sends it to the user's device, which then displays the response on a web browser so that the user can easily view it.

[0118] Specific example explanation

[0119] A specific example of the operation of the system is shown below.

[0120] 1. The user accesses the Security Concierge web interface using a terminal.

[0121] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[0122] 3. The server receives this question, and the natural language processing module extracts the keywords "ChatGPT" and "security."

[0123] 4. The server queries the database based on these keywords to find relevant information.

[0124] 5. The database returns the search results (e.g., the company's security policy regarding the use of ChatGPT).

[0125] 6. The server uses a natural language generation module to generate a response that states, "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted."

[0126] 7. The server formats this response in HTML format and sends it to the terminal.

[0127] 8. The device displays this response in the display area on the web browser so that the user can confirm it.

[0128] In this way, the system can provide quick and accurate answers to user questions about the use of generative AI and provide guidelines for companies to use generative AI with confidence.

[0129] Prompt Sentence Examples

[0130] Here are some examples of prompts to input to a generative AI model:

[0131] "What are the security risks when using ChatGPT in-house?"

[0132] "I want to know about the legal risks and countermeasures when introducing generative AI in-house."

[0133] "What are the best practices for enterprises using generative AI?"

[0134] Using this prompt, the system can generate an appropriate response to help the user obtain the information they need.

[0135] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0136] Step 1:

[0137] A user uses a terminal to access the web interface of the security concierge system via a web browser. This is equivalent to issuing an HTTP request and accessing the server. The user enters a question in the question form and clicks the send button. This operation causes the user input means to send the question data to the server.

[0138] Input: A query entered by a user into an input form (e.g., "I want to use ChatGPT in my company, but are there any security issues?")

[0139] Output: Query data sent from the user input means to the server

[0140] Step 2:

[0141] The server receives queries sent by users via web server software (e.g., Nginx or Apache) and a framework implemented in Python (e.g., Django or Flask) handles the processing of the queries. The server then passes the queries to a natural language processor.

[0142] Input: Query data sent from user input means

[0143] Output: Unparsed query data passed to natural language processing tools

[0144] Step 3:

[0145] The server analyzes the query using natural language processing. It uses a natural language processing module (e.g., NLTK or spaCy) to extract important keywords and contextual information from the query. Specifically, it performs processes such as tokenization, morphological analysis, and tagging. For example, it extracts the keywords "ChatGPT" and "security" from the query, "We want to use ChatGPT in our company, but are there any security issues?"

[0146] Input: Unparsed query data

[0147] Output: Extracted keywords and context information (e.g., "ChatGPT", "security")

[0148] Step 4:

[0149] The server searches for relevant information using a database connection method based on the extracted keywords. Specifically, it generates an SQL query and accesses the database (MySQL or PostgreSQL) using a database connection library (e.g., SQLAlchemy or Psycopg2). An example of the query content is "SELECT FROM security_policies WHERE keywords LIKE '%ChatGPT%' AND keywords LIKE '%security%'".

[0150] Input: Extracted keywords and context information

[0151] Output: Relevant information retrieved from the database (e.g., security policy regarding ChatGPT usage)

[0152] Step 5:

[0153] The server uses natural language generation to generate answers based on information retrieved from the database. It uses natural language generation modules (such as OpenAI's GPT-3 or GPT-4) to format the retrieved information in an easy-to-understand format.

[0154] Input: Relevant information retrieved from the database

[0155] Output: Answer generated by natural language generation (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted.")

[0156] Step 6:

[0157] The server formats the generated response into an appropriate HTML format and sends it to the user's device as an HTTP response, setting the response headers and status code appropriately.

[0158] Input: Answer generated by a natural language generator

[0159] Output: HTML formatted response data sent to the terminal as an HTTP response

[0160] Step 7:

[0161] The terminal displays the received HTML response in a web browser using an HTML rendering engine, and the content is displayed in a format that the user can visually confirm.

[0162] Input: HTML formatted response data sent from the server

[0163] Output: Answer displayed in web browser (e.g., "ChatGPT may pose some security risks, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are performed.")

[0164] As described above, the overall processing of the system is realized in a series of steps, and at each step the input data is appropriately processed and analyzed, ultimately providing the appropriate answer to the user.

[0165] (Application example 1)

[0166] Next, a description will be given of Application 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."

[0167] There is a need to manage the security and legal risks associated with the use of generative AI and provide an environment in which companies can use it with peace of mind. However, current systems often lack specific guidelines for specific risks associated with the use of generative AI and information for legal compliance. Furthermore, there are no systems in place that can assess risks and generate guidelines in a format that is easy for users to understand. This has resulted in a lack of comprehensive solutions for reducing the risks that arise when companies operate generative AI.

[0168] 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.

[0169] In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a means for generating guidelines using a generative AI model, and a display means, which makes it possible to search for information on relevant laws and guidelines based on a question from a user, generate a risk assessment and specific guidelines using the generative AI model, and display them in an easy-to-understand manner for the user.

[0170] "User input means" refers to an interface for accepting questions or requests from users, and includes smartphone applications, web browsers, and other input devices.

[0171] "Natural language processing means" is a technology for analyzing queries (questions or requests) obtained from user input means and extracting important keywords and contextual information.

[0172] The "database connection means" has a function of accessing a database to search for related information based on a query analyzed by the natural language processing means.

[0173] The "natural language generation means" is a technology for generating answers that are easy for the user to understand based on information acquired by the database connection means.

[0174] The "display means" is an interface for presenting the answer generated by the natural language generation means to the user, and includes a smartphone screen, a computer display, and other display devices.

[0175] "Laws and guidelines" includes rules, regulations, and guidelines that companies must comply with when using generative AI.

[0176] A "generative AI model" is an artificial intelligence technology that generates appropriate answers and guidelines based on information from users.

[0177] A "prompt" is a sentence or instruction that serves as a starting point for a generative AI model to generate appropriate answers or guidelines.

[0178] This invention provides a system for implementing security risk assessment and guideline generation. This system is composed of a user input means, a natural language processing means, a database connection means, a natural language generation means, a guideline generation means using a generative AI model, and a display means. Each component and its operation procedure will be specifically explained below.

[0179] 1. User Input Method

[0180] The user input means provides an interface for users to input questions or requests to the generative AI. Specifically, this corresponds to a smartphone application, a web browser, etc. When the user inputs a question, the query is sent to the system.

[0181] 2. Natural Language Processing Methods

[0182] The server is equipped with a natural language processing unit to analyze queries received from users, using technologies such as spaCy to extract important keywords and contextual information from the queries.

[0183] 3. Database connection method

[0184] Based on the extracted keywords and context information, the server searches for relevant information using a database connection method. The database stores corporate rules and legal information. For example, an SQLite database is used.

[0185] 4. Natural language generation means

[0186] The information retrieved from the database is used by a natural language generation system on the server to generate answers in a format that is easy for users to understand. This process uses generative AI models such as OpenAI GPT-3, which allows information to be provided in a format that is easy for users to understand, even if the content is technical.

[0187] 5. Guideline generation method using generative AI model

[0188] Furthermore, it includes a means for generating specific guidelines using a generative AI model based on the generated information. The generative AI model uses the OpenAI API, and specific guidelines for companies to use generative AI safely are generated based on a prompt. For example, a prompt might be in the format of "Based on the following information, please generate guidelines for companies to use generative AI safely: While there are some potential security risks associated with using ChatGPT, it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews."

[0189] 6. Display means

[0190] Finally, the generated answers and guidelines are returned to the user's input means and presented to the user via a smartphone application or web browser, allowing the user to review the generated information and take any necessary actions.

[0191] For example, if a user types a question like "I want to use ChatGPT in my company, but are there any security issues?", the system will follow these steps:

[0192] 1. A question is received and the keywords "ChatGPT" and "security" are extracted using natural language processing.

[0193] 2. The database is searched based on the extracted keywords to obtain relevant security policies.

[0194] 3. Specific guidelines are generated using natural language generation tools and generative AI models.

[0195] 4. Users will be shown guidelines such as, "Companies can safely use ChatGPT by meeting the following conditions: 1. Restrict access outside the company network, and 2. We recommend conducting regular security reviews."

[0196] In this way, the system of the present invention provides specific guidelines to enable companies to use generative AI with confidence, reducing security and legal risks.

[0197] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0198] Step 1:

[0199] A user accesses the system via a smartphone application or web browser and enters a question. The question entered is something like, "I want to use ChatGPT in my company, but are there any security issues?" Input: User question. Output: User query data.

[0200] Step 2:

[0201] The server receives the user query data and analyzes the query using natural language processing means. It uses a natural language processing library such as SpaCy to extract important keywords and contextual information from the query. Input: User query data. Output: Extracted keywords and contextual information.

[0202] Step 3:

[0203] The server searches for relevant information through a database connection based on the extracted keywords and context information. This involves sending queries to a database containing corporate rules and legal information. Specifically, it retrieves relevant information from an SQLite database. Input: Extracted keywords and context information. Output: Database search results.

[0204] Step 4:

[0205] The server uses natural language generation to generate user-friendly answers based on information retrieved from the database. A generative AI model such as OpenAI GPT-3 handles this process, providing specialized information in a format that is easy for users to understand. Input: Database search results. Output: Generated natural language answers.

[0206] Step 5:

[0207] The server uses a generative AI model to generate specific guidelines to further develop the generated answers. The generative AI model uses the OpenAI API, and specific guidelines are generated based on the prompt text. For example, the prompt text could be, "Based on the following information, please generate guidelines for the safe use of generative AI by the company: There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews." Input: Generated answer and prompt text. Output: Specific guidelines.

[0208] Step 6:

[0209] The server uses a display means to present the generated guidelines to the user. The generated information is displayed to the user through a smartphone application or a web browser. Input: Generated guidelines. Output: Guidelines displayed to the user.

[0210] example:

[0211] The user enters a question through the application: "I would like to use ChatGPT in my company, but are there any security issues?"

[0212] The server receives and analyzes the question, extracting the keywords "ChatGPT" and "security."

[0213] The server retrieves relevant laws and guidelines from a database based on the extracted keywords.

[0214] The server generates a response based on the acquired information using natural language generation means.

[0215] Using the answers generated by the server, a generative AI model is used to generate specific guidelines.

[0216] The server displays the generated guidelines to the user on the application.

[0217] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0218] The security concierge system incorporating the emotion engine of the present invention is a system for recognizing a user's emotion and providing a more appropriate response based on the emotion. This system includes, as its main components, a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, and an emotion engine.

[0219] System configuration

[0220] 1. User Input Method

[0221] Users access the web interface via a terminal.

[0222] This interface includes a question form and a submit button, allowing users to enter and submit questions about the generated AI.

[0223] 2. Emotion Engine

[0224] The server receives the question sent from the user input means and passes the question to the emotion engine at the same time.

[0225] The emotion engine analyzes the user's emotions from the question's sentence structure and keywords, and determines emotions such as "anxiety," "interest," and "anger."

[0226] The analyzed emotional information is reflected in the subsequent natural language processing means and natural language generation means.

[0227] 3. Natural Language Processing Methods

[0228] The server analyzes the user's question using a natural language processing module along with emotional information from the emotion engine.

[0229] This module extracts important keywords and contextual information from questions, and also takes into account the output of the emotion engine.

[0230] 4. Database connection method

[0231] The server generates and sends a query to the corresponding database based on the keywords and emotional information extracted by natural language processing.

[0232] The database contains corporate security rules and legal information and searches for and returns the appropriate information.

[0233] 5. Natural language generation means

[0234] The server generates a response in a format that is easy for the user to understand based on the information acquired from the database connection means as well as the emotion information output by the emotion engine.

[0235] A natural language generation module takes over this process, making it possible to generate answers that better reflect the user's feelings.

[0236] 6. Display means

[0237] The server formats the generated response in HTML format or similar and sends it to the user's terminal.

[0238] The terminal displays the received answers on a web interface for easy confirmation by the user.

[0239] Specific examples

[0240] 1. The user accesses the Security Concierge web interface on their device.

[0241] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[0242] 3. When the server receives the question, the emotion engine simultaneously analyzes the question and determines the user's emotion as "anxiety."

[0243] 4. The server uses a natural language processing module to extract the important keywords "ChatGPT" and "security."

[0244] 5. The server sends a query to the database based on the keywords and emotion information to search for relevant company rules and legal information.

[0245] 6. The database returns the search results to the server.

[0246] 7. The server uses a natural language generation module to generate answers that also reflect emotional information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the conditions for not needing to be concerned are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[0247] 8. The server sends the generated answer to the user's device.

[0248] 9. The device displays the received answer, and the user can decide on the next action based on the answer (e.g., setting access restrictions outside the company network).

[0249] In this way, a security concierge system combined with an emotion engine can take into account the user's emotions and provide more appropriate and personalized answers.

[0250] The processing flow will be explained below.

[0251] Step 1:

[0252] The user opens a web browser on the device and accesses the Security Concierge web interface.

[0253] Step 2:

[0254] Users enter questions about the generative AI service into a question form on the interface (e.g., "I would like to use ChatGPT in my company, but are there any security issues?").

[0255] Step 3:

[0256] After the user enters the question, he or she clicks the submit button.

[0257] Step 4:

[0258] The terminal sends the user's input to the server as an HTTP request.

[0259] Step 5:

[0260] The server receives the HTTP request and extracts the user's question.

[0261] Step 6:

[0262] The server passes the user's question to the emotion engine, which analyzes the user's emotions.

[0263] Step 7:

[0264] The emotion engine (a component within the server) analyzes the input query and determines the user's emotion from the content (e.g., "anxiety," "interest," "anger," etc.).

[0265] Step 8:

[0266] The server passes the question data to a natural language processing (NLP) module along with the emotion information determined by the emotion engine.

[0267] Step 9:

[0268] The natural language processing (NLP) module analyzes the question and extracts keywords and contextual information (e.g., "ChatGPT," "security").

[0269] Step 10:

[0270] The server generates a search query based on the extracted keywords and emotion information.

[0271] Step 11:

[0272] The server sends the generated search query to the database.

[0273] Step 12:

[0274] The database searches for relevant information (e.g., company rules and legal information) based on the received query.

[0275] Step 13:

[0276] The database returns the search results to the server.

[0277] Step 14:

[0278] The server passes the acquired information and the emotional information from the emotion engine to the natural language generation (NLG) module and requests it to generate an answer.

[0279] Step 15:

[0280] The natural language generation (NLG) module generates a user-friendly answer based on the search results and sentiment information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the conditions for not needing to be concerned are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[0281] Step 16:

[0282] The server converts the generated response into HTML format and sends it to the user's terminal.

[0283] Step 17:

[0284] The terminal displays the received HTML data in a web browser so that the user can easily check it.

[0285] Step 18:

[0286] The user reviews the displayed answers and decides on the next action based on them (e.g., setting restrictions on access outside the company network).

[0287] Example 2

[0288] Next, a description will be given of 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."

[0289] In conventional information search systems, the main purpose is to provide appropriate information in response to user queries. However, because they simply perform keyword matching without considering the user's emotions, it is difficult to provide answers that are in tune with the user's emotions. If a user asks a question harboring a specific emotion, such as anxiety or curiosity, an answer that ignores that emotion is likely to reduce satisfaction. Therefore, the present invention aims to provide a means for analyzing a user's emotions and providing a more appropriate answer based on that emotion.

[0290] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a user input means, a natural language processing means and an emotion analysis means for analyzing the query and emotion, a database connection means for searching for related information based on the analyzed query and emotion, a natural language generation means for generating an easy-to-understand answer based on the acquired information and the user's emotion, and a display means for presenting the generated answer to the user. This makes it possible to analyze the user's emotion and provide an appropriate and personalized answer based on the analysis.

[0291] "User input means" means an interface through which a user accesses the system to input and transmit information.

[0292] "Natural language processing means" refers to technology for analyzing queries entered by users and extracting important keywords and contextual information.

[0293] "Emotion analysis means" refers to a technology for analyzing emotions based on user input and determining a specific emotion (e.g., anxiety, interest, anger, etc.).

[0294] The "database connection means" means a means for transmitting a query to a database to search for related data and obtain information based on the information analyzed by the natural language processing means and the sentiment analysis means.

[0295] "Natural language generation means" refers to a technology that reflects information obtained from a database and user emotional information, and generates answers in an easy-to-understand format based on that information.

[0296] "Display means" refers to an interface for presenting the generated answers to the user, typically a web interface or a screen display.

[0297] A specific example of the invention is described below. This system is a security concierge system that analyzes a user's emotions and provides appropriate answers based on the analysis. This system is mainly composed of a user input means, a natural language processing means, an emotion analysis means, a database connection means, a natural language generation means, and a display means.

[0298] Hardware and software used

[0299] 1. Hardware

[0300] Server: A computer that processes user requests and performs the necessary calculations.

[0301] Terminal: The device from which the user accesses the site (PC, tablet, smartphone, etc.).

[0302] 2. Software

[0303] Web Interface: A UI for users to enter questions.

[0304] Natural language processing module: Software that analyzes the user's question.

[0305] Emotion engine: Software that analyzes user emotions.

[0306] Database: Data storage where company security regulations and legal information are kept.

[0307] Natural Language Generation Module: Software that generates responses.

[0308] Specific operation of the system

[0309] 1. User Input

[0310] The user accesses the system's web interface using a terminal, enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the inquiry form, and presses the submit button.

[0311] 2. User Question and Sentiment Analysis

[0312] The server receives the question data and passes it to the emotion engine, which analyzes the question and determines the emotion, such as "anxiety."

[0313] 3. Natural Language Processing

[0314] The server passes the question, including the emotion information from the emotion engine, to the natural language processing module, which extracts important keywords such as "ChatGPT" and "security" from the question.

[0315] 4. Database query generation and submission

[0316] The server generates a database query based on the extracted keywords and emotion information and sends it to the database.

[0317] 5. Information acquisition

[0318] The database searches for relevant information and returns the results to the server, including information such as terms and security guidelines for using ChatGPT safely within your company.

[0319] 6. Answer generation

[0320] The server uses a natural language generation module to generate a response to the user based on the information retrieved from the database and the output of the emotion engine. For example, it could generate a specific response such as, "There are some security risks associated with using ChatGPT, but it can be used safely if the following conditions are met: 1. Access outside the company network is restricted, and 2. Regular security reviews are required."

[0321] 7. View Answers

[0322] The server formats the generated response in HTML format and sends it to the user's device. The device displays the received response on a web interface, and the user decides what to do next based on the response.

[0323] Specific examples

[0324] A specific example is shown below.

[0325] 1. User Input Example

[0326] "We want to use ChatGPT in our company, but are there any security issues?"

[0327] 2. Example answers generated

[0328] "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met to eliminate them. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews must be conducted."

[0329] The above is an embodiment of the security concierge system incorporating an emotion engine.

[0330] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0331] Step 1: User enters and submits question

[0332] Input: The user accesses the web interface using a terminal, enters the question "I would like to use ChatGPT in my company, but are there any security issues?" and presses the send button.

[0333] Processing: The terminal sends the entered question data to the server.

[0334] Output: The question data is sent to the server.

[0335] Step 2: The server receives the question and analyzes the sentiment

[0336] Input: The server receives the query data sent from the terminal.

[0337] Processing: The server passes the question data to the emotion engine, which analyzes the user's emotions. The emotion engine performs natural language analysis based on the text information and determines emotions such as "anxiety."

[0338] Output: Emotional information (e.g., "anxiety") is generated as the analysis result.

[0339] Step 3: The server passes the question to the natural language processing module for analysis.

[0340] Input: The server passes the question data including the emotion information obtained from the emotion engine to the natural language processing module.

[0341] Processing: The natural language processing module analyzes the question data and extracts important keywords and contextual information (e.g., "ChatGPT" and "security").

[0342] Output: Extracted keywords and context information.

[0343] Step 4: Server generates database query

[0344] Input: The server generates a query to the database based on keywords and sentiment information obtained from the natural language processing module.

[0345] Processing: Generate queries in a specified format (e.g., "ChatGPT usage security risk") based on keywords (e.g., "ChatGPT" and "security") and sentiment information (e.g., "anxiety").

[0346] Output: The generated database query.

[0347] Step 5: Server sends query to database

[0348] Input: The server receives the generated database query.

[0349] Processing: The server sends a database query to the database through the database connection means.

[0350] Output: The database query is passed to the database.

[0351] Step 6: The database searches for the corresponding information and returns it to the server

[0352] Input: The database retrieves information based on the received query.

[0353] Processing: The database searches for information that matches the query and returns the results to the server, such as terms and security guidelines for using ChatGPT safely within your company.

[0354] Output: Search result information.

[0355] Step 7: Server Generates Answer

[0356] Input: The server receives the information retrieved from the database and the emotion information.

[0357] Processing: The server uses a natural language generation module to generate a specific response to the user based on the acquired information and sentiment information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met that eliminate the need for concern. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[0358] Output: The generated answer.

[0359] Step 8: Present the server-generated answer to the user

[0360] Input: The server receives the generated answer.

[0361] Processing: The server formats the response in HTML format or similar and sends it to the user's terminal.

[0362] Output: The formatted response data is sent to the terminal.

[0363] Step 9: Your device will display your answer

[0364] Input: The terminal receives the response data sent from the server.

[0365] Processing: The terminal displays the received response on the web interface.

[0366] Output: The user can check the answer on the display screen.

[0367] (Application example 2)

[0368] Next, a description will be given of Application 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."

[0369] Conventional security concierge systems can only generate simple answers to user input, making it difficult to provide personalized answers that take the user's emotions into account. This has led to the problem that they are unable to provide a highly satisfying service, especially to users who have concerns or questions.

[0370] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, and an emotion engine. This makes it possible to provide a personalized answer that takes into account the user's emotions.

[0371] The "user input means" is an interface for a user to input a query.

[0372] "Natural language processing means" is a means for analyzing a user's query and extracting keywords and contextual information.

[0373] The "database connection means" is a means for searching and acquiring related information from a database based on keywords extracted by the natural language processing means.

[0374] The "natural language generation means" is a means for generating a response in a format that is easy for the user to understand, based on information acquired from the database connection means.

[0375] The "display means" is a device or system for presenting the generated answer to the user.

[0376] An "emotion engine" is a system that recognizes emotions contained in user input and provides that emotion information to other processing means.

[0377] The present invention provides a security concierge system that provides personalized answers taking into account the emotions of a user. The system includes a user input unit, an emotion engine, a natural language processing unit, a database connection unit, a natural language generation unit, and a display unit.

[0378] System configuration

[0379] 1. User Input Method

[0380] Users enter queries through a smartphone application or web interface, which also includes auto-completion and voice input.

[0381] 2. Emotion Engine

[0382] The server passes the query received from the user input means to the emotion engine. The emotion engine analyzes the user's input text and voice data and classifies the user's emotions into categories such as "anxiety," "anger," and "interest." This analysis uses an emotion analysis API such as IBM Watson (registered trademark) Tone Analyzer.

[0383] 3. Natural Language Processing Methods

[0384] The server analyzes the user's query using a natural language processing module along with emotional information from the emotion engine. It uses the Google® Natural Language API to extract key keywords and contextual information.

[0385] 4. Database connection method

[0386] Based on the keywords and sentiment information extracted through natural language processing, the server queries an internal security database, which contains information on the company's security regulations and laws, to retrieve relevant information.

[0387] 5. Natural language generation means

[0388] The server generates answers in a user-friendly format using OpenAI's GPT-3 or similar software based on the information obtained from the database connection means and the output of the emotion engine. Generation based on emotion information enables more personalized answers.

[0389] 6. Display means

[0390] The server formats the generated answer in HTML format or similar and sends it to the user's device, which then displays the received answer on a web interface or application UI.

[0391] Specific examples

[0392] If a user enters the query, "I want to use ChatGPT in my company, but are there any security issues?", the system will behave as follows:

[0393] 1. Upon receiving the query, the server sends a question to the emotion engine and determines the user's emotion as "anxiety."

[0394] 2. The server uses a natural language processing module to extract important keywords such as "ChatGPT" and "security."

[0395] 3. The server sends a query to the database based on the extracted keywords and emotion information to obtain relevant company rules and legal information.

[0396] 4. The server generates a response using OpenAI's GPT-3 based on the acquired information and emotion information. The response generated is, "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required."

[0397] 5. The server sends the generated answer to the user's smartphone and displays it on the application.

[0398] Prompt Sentence Examples

[0399] The user is feeling anxious. Please provide gentle advice based on the following information:

[0400] 1. General risks in implementing new security measures

[0401] 2. Regular security reviews are necessary

[0402] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0403] Step 1:

[0404] The user opens the application on their smartphone and inputs a question. The input question is sent to the device as text or voice. The input data includes specific questions such as, "We want to use ChatGPT in our company, but are there any security issues?" This data is sent to the server, which initiates the next processing step.

[0405] Step 2:

[0406] The server passes the query received from the user's input means to the emotion engine. The emotion engine analyzes the text data of the received query and classifies the user's emotion. Specifically, it uses IBM Watson's Tone Analyzer to determine emotions such as "anxiety," "anger," and "interest" from the input question. Emotional information such as "anxiety" is output as the analysis result.

[0407] Step 3:

[0408] The server then sends the query, along with the emotion information from the emotion engine, to the natural language processing module, which uses the Google Natural Language API to analyze the query and extract important keywords and contextual information. For example, keywords like "ChatGPT" and "security" are extracted and sent to the next processing step.

[0409] Step 4:

[0410] The server uses the extracted keywords and emotion information to search for relevant information using a database connection method. The database stores information on corporate security rules and laws. An SQL query is sent to this database to retrieve the appropriate information. For example, information on "security risks regarding the internal use of ChatGPT" is returned as a search result.

[0411] Step 5:

[0412] The server uses natural language generation to generate responses to users based on the information and emotional information obtained from the database connection means. It uses OpenAI's GPT-3 model to create prompts based on the input data and generate emotional responses. A specific example of output might be something like, "There are some risks associated with using ChatGPT, but it can be used if appropriate measures are taken."

[0413] Step 6:

[0414] The server formats the generated answer in HTML format and sends it to the user's device. The device displays the received answer on the application. The user can then decide what to do next based on the answer.

[0415] Specific actions

[0416] Users use their smartphones to enter and submit questions.

[0417] The server uses an emotion engine to analyze the user's emotions.

[0418] The server uses a natural language processing module to extract important keywords.

[0419] The server queries the database to retrieve the appropriate information.

[0420] The server uses natural language generation means to generate answers that are easy for the user to understand.

[0421] The server sends the generated response to the terminal, which displays it.

[0422] 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 audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio 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 audio data.

[0423] 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> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with 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.

[0424] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0425] [Second embodiment]

[0426] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0427] 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.

[0428] 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 WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0429] 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.

[0430] 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 audio in accordance with instructions from the processor 46.

[0431] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and 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 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0432] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0433] Fig. 4 shows an example of the 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.

[0434] 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.

[0435] 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.

[0436] 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.

[0437] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0438] The security concierge system of the present invention is a system for reducing security and legal risks associated with the use of generative AI in companies, allowing companies to use generative AI with peace of mind. This system includes, as its main components, a user input means, a natural language processing means, a database connection means, a natural language generation means, and a display means.

[0439] System configuration

[0440] 1. User Input Method

[0441] Users access the web interface via a terminal.

[0442] This interface includes a question form and a submit button, allowing users to enter and submit questions about the generated AI.

[0443] 2. Natural Language Processing Methods

[0444] The server receives a question sent from a user input means.

[0445] The received question is analyzed by a natural language processing module to extract the most important keywords and context.

[0446] 3. Database connection method

[0447] The server uses the keywords extracted by natural language processing to send queries to a database to search for relevant information.

[0448] This database stores information about corporate security rules and regulations.

[0449] 4. Natural language generation means

[0450] The server generates an answer that is easy for the user to understand based on the information retrieved from the database.

[0451] A natural language generation module takes over this process to generate the most appropriate answer to the query.

[0452] 5. Display means

[0453] The server sends the generated answer to the user's terminal.

[0454] The terminal displays the received answers on a web interface for easy confirmation by the user.

[0455] Specific examples

[0456] 1. The user accesses the Security Concierge web interface on their device.

[0457] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[0458] 3. The server receives the question, and the natural language processing module extracts the key keywords "ChatGPT" and "security."

[0459] 4. Based on these keywords, the server searches the database for relevant company rules and legal information.

[0460] 5. The database returns the search results (e.g., the company's security policy regarding the use of ChatGPT).

[0461] 6. The server uses a natural language generation module to generate a response that states, "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, and 2. Regular security reviews are conducted."

[0462] 7. The server formats this response in HTML or other format and sends it to the terminal.

[0463] 8. The terminal displays the received answer to the user, who can then take necessary action based on the answer.

[0464] In this way, the system can quickly and accurately answer user questions about the use of generative AI and provide guidelines for companies to use generative AI with confidence.

[0465] The processing flow will be explained below.

[0466] Step 1:

[0467] The user opens a web browser on the device and accesses the Security Concierge web interface.

[0468] Step 2:

[0469] Users enter questions about the generative AI service into a question form on the interface (e.g., "I would like to use ChatGPT in my company, but are there any security issues?").

[0470] Step 3:

[0471] After the user enters the question, he or she clicks the submit button.

[0472] Step 4:

[0473] The terminal sends the user's input to the server as an HTTP request.

[0474] Step 5:

[0475] The server receives the HTTP request and extracts the user's question.

[0476] Step 6:

[0477] The server passes the question to a natural language processing (NLP) module to begin parsing it.

[0478] Step 7:

[0479] The natural language processing (NLP) module analyzes the question and extracts keywords and contextual information (e.g., "ChatGPT," "security").

[0480] Step 8:

[0481] The server generates a search query based on the extracted keywords.

[0482] Step 9:

[0483] The server sends the generated search query to the database.

[0484] Step 10:

[0485] The database searches for relevant company rules and legal information based on the received query.

[0486] Step 11:

[0487] The database returns the search results to the server.

[0488] Step 12:

[0489] The server passes the retrieved information to a natural language generation (NLG) module to begin generating an answer.

[0490] Step 13:

[0491] The Natural Language Generation (NLG) module generates answers based on the search results in a user-friendly format (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted").

[0492] Step 14:

[0493] The server converts the generated response into HTML format and sends it to the user's terminal.

[0494] Step 15:

[0495] The terminal displays the received HTML data in a web browser.

[0496] Step 16:

[0497] The user confirms the answers on the screen and decides on the next action (e.g., setting restrictions on access outside the company network).

[0498] Example 1

[0499] Next, a description will be given of 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."

[0500] The use of generative AI in companies involves security and legal risks, making it difficult to use safely. This can delay the introduction of generative AI or cause problems after its introduction. The objective of this invention is to provide a system for reducing these risks and operating generative AI safely and efficiently.

[0501] 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.

[0502] In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, a means for the user input means to receive a user input via a web interface and send a query by operating a send button, a means for the natural language processing means to analyze keywords and context information in the query using a natural language processing module, a means for the natural language generation means to generate an answer based on information obtained from a database using the natural language generation module, and a means for the display means to receive the generated answer and display it on the web interface. This reduces security and legal risks when using generative AI, allowing companies to use generative AI with peace of mind.

[0503] "User input means" refers to a means by which a user provides input to the system.

[0504] "Natural language processing means" is a means for analyzing queries received from users and extracting important keywords and contextual information.

[0505] The "database connection means" is a means for accessing a database based on a query and searching for related information.

[0506] "Natural language generation means" is a means for generating appropriate answers to users based on acquired information.

[0507] The "display means" is a means for presenting the generated answer to the user.

[0508] A "web interface" is an interface through which a user interacts with a system via a web browser.

[0509] A "natural language processing module" is a software module for analyzing input text and extracting keywords and contextual information.

[0510] A "natural language generation module" is a software module for generating answers in natural language based on analyzed information.

[0511] A "server" is a computer that performs the central processing of the system, and is a device that receives input from users and analyzes, processes, and generates responses.

[0512] A "query" refers to a question or request made by a user to a system.

[0513] The security concierge system of the present invention is intended to enable the safe use of generative AI in companies and to reduce security and legal risks. This system is implemented using the following hardware and software.

[0514] User Input Method

[0515] A user accesses the web interface of the security concierge system using a web browser. This interface includes a question form and a submit button that serves to transmit the query entered by the user to the server.

[0516] Natural language processing tools

[0517] The server receives queries sent by users. This processing is done using web server software such as Nginx or Apache. The received queries are processed using web frameworks implemented in Python, such as Django or Flask. The server then calls a natural language processing module to analyze the queries. This module uses natural language processing libraries such as NLTK or spaCy.

[0518] Database connection method

[0519] The server uses the keywords and contextual information extracted from the query to send an SQL query to a database. This database stores information about corporate security rules and regulations, and is typically an RDBMS such as MySQL or PostgreSQL. The server accesses the database using a database connection library (e.g., SQLAlchemy or Psycopg2) to retrieve the relevant information.

[0520] Natural language generation means

[0521] The server uses a natural language generation module based on the information retrieved from the database to generate answers for the user. This module uses OpenAI's GPT-3 or GPT-4, for example. Answers are generated in a format that is easy for the user to understand.

[0522] Display means

[0523] The server formats the generated response in HTML format or similar and sends it to the user's device, which then displays the response on a web browser so that the user can easily view it.

[0524] Specific example explanation

[0525] A specific example of the operation of the system is shown below.

[0526] 1. The user accesses the Security Concierge web interface using a terminal.

[0527] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[0528] 3. The server receives this question, and the natural language processing module extracts the keywords "ChatGPT" and "security."

[0529] 4. The server queries the database based on these keywords to find relevant information.

[0530] 5. The database returns the search results (e.g., the company's security policy regarding the use of ChatGPT).

[0531] 6. The server uses a natural language generation module to generate a response that states, "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted."

[0532] 7. The server formats this response in HTML format and sends it to the terminal.

[0533] 8. The device displays this response in the display area on the web browser so that the user can confirm it.

[0534] In this way, the system can provide quick and accurate answers to user questions about the use of generative AI and provide guidelines for companies to use generative AI with confidence.

[0535] Prompt Sentence Examples

[0536] Here are some examples of prompts to input to a generative AI model:

[0537] "What are the security risks when using ChatGPT in-house?"

[0538] "I want to know about the legal risks and countermeasures when introducing generative AI in-house."

[0539] "What are the best practices for enterprises using generative AI?"

[0540] Using this prompt, the system can generate an appropriate response to help the user obtain the information they need.

[0541] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0542] Step 1:

[0543] A user uses a terminal to access the web interface of the security concierge system via a web browser. This is equivalent to issuing an HTTP request and accessing the server. The user enters a question in the question form and clicks the send button. This operation causes the user input means to send the question data to the server.

[0544] Input: A query entered by a user into an input form (e.g., "I want to use ChatGPT in my company, but are there any security issues?")

[0545] Output: Query data sent from the user input means to the server

[0546] Step 2:

[0547] The server receives queries sent by users via web server software (e.g., Nginx or Apache) and a framework implemented in Python (e.g., Django or Flask) handles the processing of the queries. The server then passes the queries to a natural language processor.

[0548] Input: Query data sent from user input means

[0549] Output: Unparsed query data passed to natural language processing tools

[0550] Step 3:

[0551] The server analyzes the query using natural language processing. It uses a natural language processing module (e.g., NLTK or spaCy) to extract important keywords and contextual information from the query. Specifically, it performs processes such as tokenization, morphological analysis, and tagging. For example, it extracts the keywords "ChatGPT" and "security" from the query, "We want to use ChatGPT in our company, but are there any security issues?"

[0552] Input: Unparsed query data

[0553] Output: Extracted keywords and context information (e.g., "ChatGPT", "security")

[0554] Step 4:

[0555] The server searches for relevant information using a database connection method based on the extracted keywords. Specifically, it generates an SQL query and accesses the database (MySQL or PostgreSQL) using a database connection library (e.g., SQLAlchemy or Psycopg2). An example of the query content is "SELECT FROM security_policies WHERE keywords LIKE '%ChatGPT%' AND keywords LIKE '%security%'".

[0556] Input: Extracted keywords and context information

[0557] Output: Relevant information retrieved from the database (e.g., security policy regarding ChatGPT usage)

[0558] Step 5:

[0559] The server uses natural language generation to generate answers based on information retrieved from the database. It uses natural language generation modules (such as OpenAI's GPT-3 or GPT-4) to format the retrieved information in an easy-to-understand format.

[0560] Input: Relevant information retrieved from the database

[0561] Output: Answer generated by natural language generation (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted.")

[0562] Step 6:

[0563] The server formats the generated response into an appropriate HTML format and sends it to the user's device as an HTTP response, setting the response headers and status code appropriately.

[0564] Input: Answer generated by a natural language generator

[0565] Output: HTML formatted response data sent to the terminal as an HTTP response

[0566] Step 7:

[0567] The terminal displays the received HTML response in a web browser using an HTML rendering engine, and the content is displayed in a format that the user can visually confirm.

[0568] Input: HTML formatted response data sent from the server

[0569] Output: Answer displayed in web browser (e.g., "ChatGPT may pose some security risks, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are performed.")

[0570] As described above, the overall processing of the system is realized in a series of steps, and at each step the input data is appropriately processed and analyzed, ultimately providing the appropriate answer to the user.

[0571] (Application example 1)

[0572] Next, a description will be given of Application 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."

[0573] There is a need to manage the security and legal risks associated with the use of generative AI and provide an environment in which companies can use it with peace of mind. However, current systems often lack specific guidelines for specific risks associated with the use of generative AI and information for legal compliance. Furthermore, there are no systems in place that can assess risks and generate guidelines in a format that is easy for users to understand. This has resulted in a lack of comprehensive solutions for reducing the risks that arise when companies operate generative AI.

[0574] 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.

[0575] In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a means for generating guidelines using a generative AI model, and a display means, which makes it possible to search for information on relevant laws and guidelines based on a question from a user, generate a risk assessment and specific guidelines using the generative AI model, and display them in an easy-to-understand manner for the user.

[0576] "User input means" refers to an interface for accepting questions or requests from users, and includes smartphone applications, web browsers, and other input devices.

[0577] "Natural language processing means" is a technology for analyzing queries (questions or requests) obtained from user input means and extracting important keywords and contextual information.

[0578] The "database connection means" has a function of accessing a database to search for related information based on a query analyzed by the natural language processing means.

[0579] The "natural language generation means" is a technology for generating answers that are easy for the user to understand based on information acquired by the database connection means.

[0580] The "display means" is an interface for presenting the answer generated by the natural language generation means to the user, and includes a smartphone screen, a computer display, and other display devices.

[0581] "Laws and guidelines" includes rules, regulations, and guidelines that companies must comply with when using generative AI.

[0582] A "generative AI model" is an artificial intelligence technology that generates appropriate answers and guidelines based on information from users.

[0583] A "prompt" is a sentence or instruction that serves as a starting point for a generative AI model to generate appropriate answers or guidelines.

[0584] This invention provides a system for implementing security risk assessment and guideline generation. This system is composed of a user input means, a natural language processing means, a database connection means, a natural language generation means, a guideline generation means using a generative AI model, and a display means. Each component and its operation procedure will be specifically explained below.

[0585] 1. User Input Method

[0586] The user input means provides an interface for users to input questions or requests to the generative AI. Specifically, this corresponds to a smartphone application, a web browser, etc. When the user inputs a question, the query is sent to the system.

[0587] 2. Natural Language Processing Methods

[0588] The server is equipped with a natural language processing unit to analyze queries received from users, using technologies such as spaCy to extract important keywords and contextual information from the queries.

[0589] 3. Database connection method

[0590] Based on the extracted keywords and context information, the server searches for relevant information using a database connection method. The database stores corporate rules and legal information. For example, an SQLite database is used.

[0591] 4. Natural language generation means

[0592] The information retrieved from the database is used by a natural language generation system on the server to generate answers in a format that is easy for users to understand. This process uses generative AI models such as OpenAI GPT-3, which allows information to be provided in a format that is easy for users to understand, even if the content is technical.

[0593] 5. Guideline generation method using generative AI model

[0594] Furthermore, it includes a means for generating specific guidelines using a generative AI model based on the generated information. The generative AI model uses the OpenAI API, and specific guidelines for companies to use generative AI safely are generated based on a prompt. For example, a prompt might be in the format of "Based on the following information, please generate guidelines for companies to use generative AI safely: While there are some potential security risks associated with using ChatGPT, it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews."

[0595] 6. Display means

[0596] Finally, the generated answers and guidelines are returned to the user's input means and presented to the user via a smartphone application or web browser, allowing the user to review the generated information and take any necessary actions.

[0597] For example, if a user types a question like "I want to use ChatGPT in my company, but are there any security issues?", the system will follow these steps:

[0598] 1. A question is received and the keywords "ChatGPT" and "security" are extracted using natural language processing.

[0599] 2. The database is searched based on the extracted keywords to obtain relevant security policies.

[0600] 3. Specific guidelines are generated using natural language generation tools and generative AI models.

[0601] 4. Users will be shown guidelines such as, "Companies can safely use ChatGPT by meeting the following conditions: 1. Restrict access outside the company network, and 2. We recommend conducting regular security reviews."

[0602] In this way, the system of the present invention provides specific guidelines to enable companies to use generative AI with confidence, reducing security and legal risks.

[0603] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0604] Step 1:

[0605] A user accesses the system via a smartphone application or web browser and enters a question. The question entered is something like, "I want to use ChatGPT in my company, but are there any security issues?" Input: User question. Output: User query data.

[0606] Step 2:

[0607] The server receives the user query data and analyzes the query using natural language processing means. It uses a natural language processing library such as SpaCy to extract important keywords and contextual information from the query. Input: User query data. Output: Extracted keywords and contextual information.

[0608] Step 3:

[0609] The server searches for relevant information through a database connection based on the extracted keywords and context information. This involves sending queries to a database containing corporate rules and legal information. Specifically, it retrieves relevant information from an SQLite database. Input: Extracted keywords and context information. Output: Database search results.

[0610] Step 4:

[0611] The server uses natural language generation to generate user-friendly answers based on information retrieved from the database. A generative AI model such as OpenAI GPT-3 handles this process, providing specialized information in a format that is easy for users to understand. Input: Database search results. Output: Generated natural language answers.

[0612] Step 5:

[0613] The server uses a generative AI model to generate specific guidelines to further develop the generated answers. The generative AI model uses the OpenAI API, and specific guidelines are generated based on the prompt text. For example, the prompt text could be, "Based on the following information, please generate guidelines for the safe use of generative AI by the company: There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews." Input: Generated answer and prompt text. Output: Specific guidelines.

[0614] Step 6:

[0615] The server uses a display means to present the generated guidelines to the user. The generated information is displayed to the user through a smartphone application or a web browser. Input: Generated guidelines. Output: Guidelines displayed to the user.

[0616] example:

[0617] The user enters a question through the application: "I would like to use ChatGPT in my company, but are there any security issues?"

[0618] The server receives and analyzes the question, extracting the keywords "ChatGPT" and "security."

[0619] The server retrieves relevant laws and guidelines from a database based on the extracted keywords.

[0620] The server generates a response based on the acquired information using natural language generation means.

[0621] Using the answers generated by the server, a generative AI model is used to generate specific guidelines.

[0622] The server displays the generated guidelines to the user on the application.

[0623] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0624] The security concierge system incorporating the emotion engine of the present invention is a system for recognizing a user's emotion and providing a more appropriate response based on the emotion. This system includes, as its main components, a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, and an emotion engine.

[0625] System configuration

[0626] 1. User Input Method

[0627] Users access the web interface via a terminal.

[0628] This interface includes a question form and a submit button, allowing users to enter and submit questions about the generated AI.

[0629] 2. Emotion Engine

[0630] The server receives the question sent from the user input means and passes the question to the emotion engine at the same time.

[0631] The emotion engine analyzes the user's emotions from the question's sentence structure and keywords, and determines emotions such as "anxiety," "interest," and "anger."

[0632] The analyzed emotional information is reflected in the subsequent natural language processing means and natural language generation means.

[0633] 3. Natural Language Processing Methods

[0634] The server analyzes the user's question using a natural language processing module along with emotional information from the emotion engine.

[0635] This module extracts important keywords and contextual information from questions, and also takes into account the output of the emotion engine.

[0636] 4. Database connection method

[0637] The server generates and sends a query to the corresponding database based on the keywords and emotional information extracted by natural language processing.

[0638] The database contains corporate security rules and legal information and searches for and returns the appropriate information.

[0639] 5. Natural language generation means

[0640] The server generates a response in a format that is easy for the user to understand based on the information acquired from the database connection means as well as the emotion information output by the emotion engine.

[0641] A natural language generation module takes over this process, making it possible to generate answers that better reflect the user's feelings.

[0642] 6. Display means

[0643] The server formats the generated response in HTML format or similar and sends it to the user's terminal.

[0644] The terminal displays the received answers on a web interface for easy confirmation by the user.

[0645] Specific examples

[0646] 1. The user accesses the Security Concierge web interface on their device.

[0647] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[0648] 3. When the server receives the question, the emotion engine simultaneously analyzes the question and determines the user's emotion as "anxiety."

[0649] 4. The server uses a natural language processing module to extract the important keywords "ChatGPT" and "security."

[0650] 5. The server sends a query to the database based on the keywords and emotion information to search for relevant company rules and legal information.

[0651] 6. The database returns the search results to the server.

[0652] 7. The server uses a natural language generation module to generate answers that also reflect emotional information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the conditions for not needing to be concerned are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[0653] 8. The server sends the generated answer to the user's device.

[0654] 9. The device displays the received answer, and the user can decide on the next action based on the answer (e.g., setting access restrictions outside the company network).

[0655] In this way, a security concierge system combined with an emotion engine can take into account the user's emotions and provide more appropriate and personalized answers.

[0656] The processing flow will be explained below.

[0657] Step 1:

[0658] The user opens a web browser on the device and accesses the Security Concierge web interface.

[0659] Step 2:

[0660] Users enter questions about the generative AI service into a question form on the interface (e.g., "I would like to use ChatGPT in my company, but are there any security issues?").

[0661] Step 3:

[0662] After the user enters the question, he or she clicks the submit button.

[0663] Step 4:

[0664] The terminal sends the user's input to the server as an HTTP request.

[0665] Step 5:

[0666] The server receives the HTTP request and extracts the user's question.

[0667] Step 6:

[0668] The server passes the user's question to the emotion engine, which analyzes the user's emotions.

[0669] Step 7:

[0670] The emotion engine (a component within the server) analyzes the input query and determines the user's emotion from the content (e.g., "anxiety," "interest," "anger," etc.).

[0671] Step 8:

[0672] The server passes the question data to a natural language processing (NLP) module along with the emotion information determined by the emotion engine.

[0673] Step 9:

[0674] The natural language processing (NLP) module analyzes the question and extracts keywords and contextual information (e.g., "ChatGPT," "security").

[0675] Step 10:

[0676] The server generates a search query based on the extracted keywords and emotion information.

[0677] Step 11:

[0678] The server sends the generated search query to the database.

[0679] Step 12:

[0680] The database searches for relevant information (e.g., company rules and legal information) based on the received query.

[0681] Step 13:

[0682] The database returns the search results to the server.

[0683] Step 14:

[0684] The server passes the acquired information and the emotional information from the emotion engine to the natural language generation (NLG) module and requests it to generate an answer.

[0685] Step 15:

[0686] The natural language generation (NLG) module generates a user-friendly answer based on the search results and sentiment information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the conditions for not needing to be concerned are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[0687] Step 16:

[0688] The server converts the generated response into HTML format and sends it to the user's terminal.

[0689] Step 17:

[0690] The terminal displays the received HTML data in a web browser so that the user can easily check it.

[0691] Step 18:

[0692] The user reviews the displayed answers and decides on the next action based on them (e.g., setting restrictions on access outside the company network).

[0693] Example 2

[0694] Next, a description will be given of 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."

[0695] In conventional information search systems, the main purpose is to provide appropriate information in response to user queries. However, because they simply perform keyword matching without considering the user's emotions, it is difficult to provide answers that are in tune with the user's emotions. If a user asks a question harboring a specific emotion, such as anxiety or curiosity, an answer that ignores that emotion is likely to reduce satisfaction. Therefore, the present invention aims to provide a means for analyzing a user's emotions and providing a more appropriate answer based on that emotion.

[0696] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a user input means, a natural language processing means and an emotion analysis means for analyzing the query and emotion, a database connection means for searching for related information based on the analyzed query and emotion, a natural language generation means for generating an easy-to-understand answer based on the acquired information and the user's emotion, and a display means for presenting the generated answer to the user. This makes it possible to analyze the user's emotion and provide an appropriate and personalized answer based on the analysis.

[0697] "User input means" means an interface through which a user accesses the system to input and transmit information.

[0698] "Natural language processing means" refers to technology for analyzing queries entered by users and extracting important keywords and contextual information.

[0699] "Emotion analysis means" refers to a technology for analyzing emotions based on user input and determining a specific emotion (e.g., anxiety, interest, anger, etc.).

[0700] The "database connection means" means a means for transmitting a query to a database to search for related data and obtain information based on the information analyzed by the natural language processing means and the sentiment analysis means.

[0701] "Natural language generation means" refers to a technology that reflects information obtained from a database and user emotional information, and generates answers in an easy-to-understand format based on that information.

[0702] "Display means" refers to an interface for presenting the generated answers to the user, typically a web interface or a screen display.

[0703] A specific example of the invention is described below. This system is a security concierge system that analyzes a user's emotions and provides appropriate answers based on the analysis. This system is mainly composed of a user input means, a natural language processing means, an emotion analysis means, a database connection means, a natural language generation means, and a display means.

[0704] Hardware and software used

[0705] 1. Hardware

[0706] Server: A computer that processes user requests and performs the necessary calculations.

[0707] Terminal: The device from which the user accesses the site (PC, tablet, smartphone, etc.).

[0708] 2. Software

[0709] Web Interface: A UI for users to enter questions.

[0710] Natural language processing module: Software that analyzes the user's question.

[0711] Emotion engine: Software that analyzes user emotions.

[0712] Database: Data storage where company security regulations and legal information are kept.

[0713] Natural Language Generation Module: Software that generates responses.

[0714] Specific operation of the system

[0715] 1. User Input

[0716] The user accesses the system's web interface using a terminal, enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the inquiry form, and presses the submit button.

[0717] 2. User Question and Sentiment Analysis

[0718] The server receives the question data and passes it to the emotion engine, which analyzes the question and determines the emotion, such as "anxiety."

[0719] 3. Natural Language Processing

[0720] The server passes the question, including the emotion information from the emotion engine, to the natural language processing module, which extracts important keywords such as "ChatGPT" and "security" from the question.

[0721] 4. Database query generation and submission

[0722] The server generates a database query based on the extracted keywords and emotion information and sends it to the database.

[0723] 5. Information acquisition

[0724] The database searches for relevant information and returns the results to the server, including information such as terms and security guidelines for using ChatGPT safely within your company.

[0725] 6. Answer generation

[0726] The server uses a natural language generation module to generate a response to the user based on the information retrieved from the database and the output of the emotion engine. For example, it could generate a specific response such as, "There are some security risks associated with using ChatGPT, but it can be used safely if the following conditions are met: 1. Access outside the company network is restricted, and 2. Regular security reviews are required."

[0727] 7. View Answers

[0728] The server formats the generated response in HTML format and sends it to the user's device. The device displays the received response on a web interface, and the user decides what to do next based on the response.

[0729] Specific examples

[0730] A specific example is shown below.

[0731] 1. User Input Example

[0732] "We want to use ChatGPT in our company, but are there any security issues?"

[0733] 2. Example answers generated

[0734] "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met to eliminate them. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews must be conducted."

[0735] The above is an embodiment of the security concierge system incorporating an emotion engine.

[0736] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0737] Step 1: User enters and submits question

[0738] Input: The user accesses the web interface using a terminal, enters the question "I would like to use ChatGPT in my company, but are there any security issues?" and presses the send button.

[0739] Processing: The terminal sends the entered question data to the server.

[0740] Output: The question data is sent to the server.

[0741] Step 2: The server receives the question and analyzes the sentiment

[0742] Input: The server receives the query data sent from the terminal.

[0743] Processing: The server passes the question data to the emotion engine, which analyzes the user's emotions. The emotion engine performs natural language analysis based on the text information and determines emotions such as "anxiety."

[0744] Output: Emotional information (e.g., "anxiety") is generated as the analysis result.

[0745] Step 3: The server passes the question to the natural language processing module for analysis.

[0746] Input: The server passes the question data including the emotion information obtained from the emotion engine to the natural language processing module.

[0747] Processing: The natural language processing module analyzes the question data and extracts important keywords and contextual information (e.g., "ChatGPT" and "security").

[0748] Output: Extracted keywords and context information.

[0749] Step 4: Server generates database query

[0750] Input: The server generates a query to the database based on keywords and sentiment information obtained from the natural language processing module.

[0751] Processing: Generate queries in a specified format (e.g., "ChatGPT usage security risk") based on keywords (e.g., "ChatGPT" and "security") and sentiment information (e.g., "anxiety").

[0752] Output: The generated database query.

[0753] Step 5: Server sends query to database

[0754] Input: The server receives the generated database query.

[0755] Processing: The server sends a database query to the database through the database connection means.

[0756] Output: The database query is passed to the database.

[0757] Step 6: The database searches for the corresponding information and returns it to the server

[0758] Input: The database retrieves information based on the received query.

[0759] Processing: The database searches for information that matches the query and returns the results to the server, such as terms and security guidelines for using ChatGPT safely within your company.

[0760] Output: Search result information.

[0761] Step 7: Server Generates Answer

[0762] Input: The server receives the information retrieved from the database and the emotion information.

[0763] Processing: The server uses a natural language generation module to generate a specific response to the user based on the acquired information and sentiment information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met that eliminate the need for concern. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[0764] Output: The generated answer.

[0765] Step 8: Present the server-generated answer to the user

[0766] Input: The server receives the generated answer.

[0767] Processing: The server formats the response in HTML format or similar and sends it to the user's terminal.

[0768] Output: The formatted response data is sent to the terminal.

[0769] Step 9: Your device will display your answer

[0770] Input: The terminal receives the response data sent from the server.

[0771] Processing: The terminal displays the received response on the web interface.

[0772] Output: The user can check the answer on the display screen.

[0773] (Application example 2)

[0774] Next, a description will be given of Application 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."

[0775] Conventional security concierge systems can only generate simple answers to user input, making it difficult to provide personalized answers that take the user's emotions into account. This has led to the problem that they are unable to provide a highly satisfying service, especially to users who have concerns or questions.

[0776] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, and an emotion engine. This makes it possible to provide a personalized answer that takes into account the user's emotions.

[0777] The "user input means" is an interface for a user to input a query.

[0778] "Natural language processing means" is a means for analyzing a user's query and extracting keywords and contextual information.

[0779] The "database connection means" is a means for searching and acquiring related information from a database based on keywords extracted by the natural language processing means.

[0780] The "natural language generation means" is a means for generating a response in a format that is easy for the user to understand, based on information acquired from the database connection means.

[0781] The "display means" is a device or system for presenting the generated answer to the user.

[0782] An "emotion engine" is a system that recognizes emotions contained in user input and provides that emotion information to other processing means.

[0783] The present invention provides a security concierge system that provides personalized answers taking into account the emotions of a user. The system includes a user input unit, an emotion engine, a natural language processing unit, a database connection unit, a natural language generation unit, and a display unit.

[0784] System configuration

[0785] 1. User Input Method

[0786] Users enter queries through a smartphone application or web interface, which also includes auto-completion and voice input.

[0787] 2. Emotion Engine

[0788] The server passes queries received from the user's input means to the emotion engine, which analyzes the user's input text and voice data and classifies the user's emotions into categories such as "anxiety," "anger," and "interest." This analysis uses emotion analysis APIs such as IBM Watson's Tone Analyzer.

[0789] 3. Natural Language Processing Methods

[0790] The server analyzes the user's query using a natural language processing module along with sentiment information from the sentiment engine, using the Google Natural Language API to extract key keywords and contextual information.

[0791] 4. Database connection method

[0792] Based on the keywords and sentiment information extracted through natural language processing, the server queries an internal security database, which contains information on the company's security regulations and laws, to retrieve relevant information.

[0793] 5. Natural language generation means

[0794] The server generates answers in a user-friendly format using OpenAI's GPT-3 or similar software based on the information obtained from the database connection means and the output of the emotion engine. Generation based on emotion information enables more personalized answers.

[0795] 6. Display means

[0796] The server formats the generated answer in HTML format or similar and sends it to the user's device, which then displays the received answer on a web interface or application UI.

[0797] Specific examples

[0798] If a user enters the query, "I want to use ChatGPT in my company, but are there any security issues?", the system will behave as follows:

[0799] 1. Upon receiving the query, the server sends a question to the emotion engine and determines the user's emotion as "anxiety."

[0800] 2. The server uses a natural language processing module to extract important keywords such as "ChatGPT" and "security."

[0801] 3. The server sends a query to the database based on the extracted keywords and emotion information to obtain relevant company rules and legal information.

[0802] 4. The server generates a response using OpenAI's GPT-3 based on the acquired information and emotion information. The response generated is, "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required."

[0803] 5. The server sends the generated answer to the user's smartphone and displays it on the application.

[0804] Prompt Sentence Examples

[0805] The user is feeling anxious. Please provide gentle advice based on the following information:

[0806] 1. General risks in implementing new security measures

[0807] 2. Regular security reviews are necessary

[0808] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0809] Step 1:

[0810] The user opens the application on their smartphone and inputs a question. The input question is sent to the device as text or voice. The input data includes specific questions such as, "We want to use ChatGPT in our company, but are there any security issues?" This data is sent to the server, which initiates the next processing step.

[0811] Step 2:

[0812] The server passes the query received from the user's input means to the emotion engine. The emotion engine analyzes the text data of the received query and classifies the user's emotion. Specifically, it uses IBM Watson's Tone Analyzer to determine emotions such as "anxiety," "anger," and "interest" from the input question. Emotional information such as "anxiety" is output as the analysis result.

[0813] Step 3:

[0814] The server then sends the query, along with the emotion information from the emotion engine, to the natural language processing module, which uses the Google Natural Language API to analyze the query and extract important keywords and contextual information. For example, keywords like "ChatGPT" and "security" are extracted and sent to the next processing step.

[0815] Step 4:

[0816] The server uses the extracted keywords and emotion information to search for relevant information using a database connection method. The database stores information on corporate security rules and laws. An SQL query is sent to this database to retrieve the appropriate information. For example, information on "security risks regarding the internal use of ChatGPT" is returned as a search result.

[0817] Step 5:

[0818] The server uses natural language generation to generate responses to users based on the information and emotional information obtained from the database connection means. It uses OpenAI's GPT-3 model to create prompts based on the input data and generate emotional responses. A specific example of output might be something like, "There are some risks associated with using ChatGPT, but it can be used if appropriate measures are taken."

[0819] Step 6:

[0820] The server formats the generated answer in HTML format and sends it to the user's device. The device displays the received answer on the application. The user can then decide what to do next based on the answer.

[0821] Specific actions

[0822] Users use their smartphones to enter and submit questions.

[0823] The server uses an emotion engine to analyze the user's emotions.

[0824] The server uses a natural language processing module to extract important keywords.

[0825] The server queries the database to retrieve the appropriate information.

[0826] The server uses natural language generation means to generate answers that are easy for the user to understand.

[0827] The server sends the generated response to the terminal, which displays it.

[0828] 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 audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio 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 audio data.

[0829] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with 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.

[0830] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0831] [Third embodiment]

[0832] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0833] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0834] 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 WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0835] 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.

[0836] 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 audio in accordance with instructions from the processor 46.

[0837] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and 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 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0838] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0839] 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.

[0840] 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.

[0841] 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.

[0842] In the headset type terminal 314, a 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.

[0843] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0844] The security concierge system of the present invention is a system for reducing security and legal risks associated with the use of generative AI in companies, allowing companies to use generative AI with peace of mind. This system includes, as its main components, a user input means, a natural language processing means, a database connection means, a natural language generation means, and a display means.

[0845] System configuration

[0846] 1. User Input Method

[0847] Users access the web interface via a terminal.

[0848] This interface includes a question form and a submit button, allowing users to enter and submit questions about the generated AI.

[0849] 2. Natural Language Processing Methods

[0850] The server receives a question sent from a user input means.

[0851] The received question is analyzed by a natural language processing module to extract the most important keywords and context.

[0852] 3. Database connection method

[0853] The server uses the keywords extracted by natural language processing to send queries to a database to search for relevant information.

[0854] This database stores information about corporate security rules and regulations.

[0855] 4. Natural language generation means

[0856] The server generates an answer that is easy for the user to understand based on the information retrieved from the database.

[0857] A natural language generation module takes over this process to generate the most appropriate answer to the query.

[0858] 5. Display means

[0859] The server sends the generated answer to the user's terminal.

[0860] The terminal displays the received answers on a web interface for easy confirmation by the user.

[0861] Specific examples

[0862] 1. The user accesses the Security Concierge web interface on their device.

[0863] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[0864] 3. The server receives the question, and the natural language processing module extracts the key keywords "ChatGPT" and "security."

[0865] 4. Based on these keywords, the server searches the database for relevant company rules and legal information.

[0866] 5. The database returns the search results (e.g., the company's security policy regarding the use of ChatGPT).

[0867] 6. The server uses a natural language generation module to generate a response that states, "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, and 2. Regular security reviews are conducted."

[0868] 7. The server formats this response in HTML or other format and sends it to the terminal.

[0869] 8. The terminal displays the received answer to the user, who can then take necessary action based on the answer.

[0870] In this way, the system can quickly and accurately answer user questions about the use of generative AI and provide guidelines for companies to use generative AI with confidence.

[0871] The processing flow will be explained below.

[0872] Step 1:

[0873] The user opens a web browser on the device and accesses the Security Concierge web interface.

[0874] Step 2:

[0875] Users enter questions about the generative AI service into a question form on the interface (e.g., "I would like to use ChatGPT in my company, but are there any security issues?").

[0876] Step 3:

[0877] After the user enters the question, he or she clicks the submit button.

[0878] Step 4:

[0879] The terminal sends the user's input to the server as an HTTP request.

[0880] Step 5:

[0881] The server receives the HTTP request and extracts the user's question.

[0882] Step 6:

[0883] The server passes the question to a natural language processing (NLP) module to begin parsing it.

[0884] Step 7:

[0885] The natural language processing (NLP) module analyzes the question and extracts keywords and contextual information (e.g., "ChatGPT," "security").

[0886] Step 8:

[0887] The server generates a search query based on the extracted keywords.

[0888] Step 9:

[0889] The server sends the generated search query to the database.

[0890] Step 10:

[0891] The database searches for relevant company rules and legal information based on the received query.

[0892] Step 11:

[0893] The database returns the search results to the server.

[0894] Step 12:

[0895] The server passes the retrieved information to a natural language generation (NLG) module to begin generating an answer.

[0896] Step 13:

[0897] The Natural Language Generation (NLG) module generates answers based on the search results in a user-friendly format (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted").

[0898] Step 14:

[0899] The server converts the generated response into HTML format and sends it to the user's terminal.

[0900] Step 15:

[0901] The terminal displays the received HTML data in a web browser.

[0902] Step 16:

[0903] The user confirms the answers on the screen and decides on the next action (e.g., setting restrictions on access outside the company network).

[0904] Example 1

[0905] Next, a description will be given of 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."

[0906] The use of generative AI in companies involves security and legal risks, making it difficult to use safely. This can delay the introduction of generative AI or cause problems after its introduction. The objective of this invention is to provide a system for reducing these risks and operating generative AI safely and efficiently.

[0907] 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.

[0908] In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, a means for the user input means to receive a user input via a web interface and send a query by operating a send button, a means for the natural language processing means to analyze keywords and context information in the query using a natural language processing module, a means for the natural language generation means to generate an answer based on information obtained from a database using the natural language generation module, and a means for the display means to receive the generated answer and display it on the web interface. This reduces security and legal risks when using generative AI, allowing companies to use generative AI with peace of mind.

[0909] "User input means" refers to a means by which a user provides input to the system.

[0910] "Natural language processing means" is a means for analyzing queries received from users and extracting important keywords and contextual information.

[0911] The "database connection means" is a means for accessing a database based on a query and searching for related information.

[0912] "Natural language generation means" is a means for generating appropriate answers to users based on acquired information.

[0913] The "display means" is a means for presenting the generated answer to the user.

[0914] A "web interface" is an interface through which a user interacts with a system via a web browser.

[0915] A "natural language processing module" is a software module for analyzing input text and extracting keywords and contextual information.

[0916] A "natural language generation module" is a software module for generating answers in natural language based on analyzed information.

[0917] A "server" is a computer that performs the central processing of the system, and is a device that receives input from users and analyzes, processes, and generates responses.

[0918] A "query" refers to a question or request made by a user to a system.

[0919] The security concierge system of the present invention is intended to enable the safe use of generative AI in companies and to reduce security and legal risks. This system is implemented using the following hardware and software.

[0920] User Input Method

[0921] A user accesses the web interface of the security concierge system using a web browser. This interface includes a question form and a submit button that serves to transmit the query entered by the user to the server.

[0922] Natural language processing tools

[0923] The server receives queries sent by users. This processing is done using web server software such as Nginx or Apache. The received queries are processed using web frameworks implemented in Python, such as Django or Flask. The server then calls a natural language processing module to analyze the queries. This module uses natural language processing libraries such as NLTK or spaCy.

[0924] Database connection method

[0925] The server uses the keywords and contextual information extracted from the query to send an SQL query to a database. This database stores information about corporate security rules and regulations, and is typically an RDBMS such as MySQL or PostgreSQL. The server accesses the database using a database connection library (e.g., SQLAlchemy or Psycopg2) to retrieve the relevant information.

[0926] Natural language generation means

[0927] The server uses a natural language generation module based on the information retrieved from the database to generate answers for the user. This module uses OpenAI's GPT-3 or GPT-4, for example. Answers are generated in a format that is easy for the user to understand.

[0928] Display means

[0929] The server formats the generated response in HTML format or similar and sends it to the user's device, which then displays the response on a web browser so that the user can easily view it.

[0930] Specific example explanation

[0931] A specific example of the operation of the system is shown below.

[0932] 1. The user accesses the Security Concierge web interface using a terminal.

[0933] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[0934] 3. The server receives this question, and the natural language processing module extracts the keywords "ChatGPT" and "security."

[0935] 4. The server queries the database based on these keywords to find relevant information.

[0936] 5. The database returns the search results (e.g., the company's security policy regarding the use of ChatGPT).

[0937] 6. The server uses a natural language generation module to generate a response that states, "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted."

[0938] 7. The server formats this response in HTML format and sends it to the terminal.

[0939] 8. The device displays this response in the display area on the web browser so that the user can confirm it.

[0940] In this way, the system can provide quick and accurate answers to user questions about the use of generative AI and provide guidelines for companies to use generative AI with confidence.

[0941] Prompt Sentence Examples

[0942] Here are some examples of prompts to input to a generative AI model:

[0943] "What are the security risks when using ChatGPT in-house?"

[0944] "I want to know about the legal risks and countermeasures when introducing generative AI in-house."

[0945] "What are the best practices for enterprises using generative AI?"

[0946] Using this prompt, the system can generate an appropriate response to help the user obtain the information they need.

[0947] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0948] Step 1:

[0949] A user uses a terminal to access the web interface of the security concierge system via a web browser. This is equivalent to issuing an HTTP request and accessing the server. The user enters a question in the question form and clicks the send button. This operation causes the user input means to send the question data to the server.

[0950] Input: A query entered by a user into an input form (e.g., "I want to use ChatGPT in my company, but are there any security issues?")

[0951] Output: Query data sent from the user input means to the server

[0952] Step 2:

[0953] The server receives queries sent by users via web server software (e.g., Nginx or Apache) and a framework implemented in Python (e.g., Django or Flask) handles the processing of the queries. The server then passes the queries to a natural language processor.

[0954] Input: Query data sent from user input means

[0955] Output: Unparsed query data passed to natural language processing tools

[0956] Step 3:

[0957] The server analyzes the query using natural language processing. It uses a natural language processing module (e.g., NLTK or spaCy) to extract important keywords and contextual information from the query. Specifically, it performs processes such as tokenization, morphological analysis, and tagging. For example, it extracts the keywords "ChatGPT" and "security" from the query, "We want to use ChatGPT in our company, but are there any security issues?"

[0958] Input: Unparsed query data

[0959] Output: Extracted keywords and context information (e.g., "ChatGPT", "security")

[0960] Step 4:

[0961] The server searches for relevant information using a database connection method based on the extracted keywords. Specifically, it generates an SQL query and accesses the database (MySQL or PostgreSQL) using a database connection library (e.g., SQLAlchemy or Psycopg2). An example of the query content is "SELECT FROM security_policies WHERE keywords LIKE '%ChatGPT%' AND keywords LIKE '%security%'".

[0962] Input: Extracted keywords and context information

[0963] Output: Relevant information retrieved from the database (e.g., security policy regarding ChatGPT usage)

[0964] Step 5:

[0965] The server uses natural language generation to generate answers based on information retrieved from the database. It uses natural language generation modules (such as OpenAI's GPT-3 or GPT-4) to format the retrieved information in an easy-to-understand format.

[0966] Input: Relevant information retrieved from the database

[0967] Output: Answer generated by natural language generation (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted.")

[0968] Step 6:

[0969] The server formats the generated response into an appropriate HTML format and sends it to the user's device as an HTTP response, setting the response headers and status code appropriately.

[0970] Input: Answer generated by a natural language generator

[0971] Output: HTML formatted response data sent to the terminal as an HTTP response

[0972] Step 7:

[0973] The terminal displays the received HTML response in a web browser using an HTML rendering engine, and the content is displayed in a format that the user can visually confirm.

[0974] Input: HTML formatted response data sent from the server

[0975] Output: Answer displayed in web browser (e.g., "ChatGPT may pose some security risks, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are performed.")

[0976] As described above, the overall processing of the system is realized in a series of steps, and at each step the input data is appropriately processed and analyzed, ultimately providing the appropriate answer to the user.

[0977] (Application example 1)

[0978] Next, a description will be given of Application 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."

[0979] There is a need to manage the security and legal risks associated with the use of generative AI and provide an environment in which companies can use it with peace of mind. However, current systems often lack specific guidelines for specific risks associated with the use of generative AI and information for legal compliance. Furthermore, there are no systems in place that can assess risks and generate guidelines in a format that is easy for users to understand. This has resulted in a lack of comprehensive solutions for reducing the risks that arise when companies operate generative AI.

[0980] 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.

[0981] In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a means for generating guidelines using a generative AI model, and a display means, which makes it possible to search for information on relevant laws and guidelines based on a question from a user, generate a risk assessment and specific guidelines using the generative AI model, and display them in an easy-to-understand manner for the user.

[0982] "User input means" refers to an interface for accepting questions or requests from users, and includes smartphone applications, web browsers, and other input devices.

[0983] "Natural language processing means" is a technology for analyzing queries (questions or requests) obtained from user input means and extracting important keywords and contextual information.

[0984] The "database connection means" has a function of accessing a database to search for related information based on a query analyzed by the natural language processing means.

[0985] The "natural language generation means" is a technology for generating answers that are easy for the user to understand based on information acquired by the database connection means.

[0986] The "display means" is an interface for presenting the answer generated by the natural language generation means to the user, and includes a smartphone screen, a computer display, and other display devices.

[0987] "Laws and guidelines" includes rules, regulations, and guidelines that companies must comply with when using generative AI.

[0988] A "generative AI model" is an artificial intelligence technology that generates appropriate answers and guidelines based on information from users.

[0989] A "prompt" is a sentence or instruction that serves as a starting point for a generative AI model to generate appropriate answers or guidelines.

[0990] This invention provides a system for implementing security risk assessment and guideline generation. This system is composed of a user input means, a natural language processing means, a database connection means, a natural language generation means, a guideline generation means using a generative AI model, and a display means. Each component and its operation procedure will be specifically explained below.

[0991] 1. User Input Method

[0992] The user input means provides an interface for users to input questions or requests to the generative AI. Specifically, this corresponds to a smartphone application, a web browser, etc. When the user inputs a question, the query is sent to the system.

[0993] 2. Natural Language Processing Methods

[0994] The server is equipped with a natural language processing unit to analyze queries received from users, using technologies such as spaCy to extract important keywords and contextual information from the queries.

[0995] 3. Database connection method

[0996] Based on the extracted keywords and context information, the server searches for relevant information using a database connection method. The database stores corporate rules and legal information. For example, an SQLite database is used.

[0997] 4. Natural language generation means

[0998] The information retrieved from the database is used by a natural language generation system on the server to generate answers in a format that is easy for users to understand. This process uses generative AI models such as OpenAI GPT-3, which allows information to be provided in a format that is easy for users to understand, even if the content is technical.

[0999] 5. Guideline generation method using generative AI model

[1000] Furthermore, it includes a means for generating specific guidelines using a generative AI model based on the generated information. The generative AI model uses the OpenAI API, and specific guidelines for companies to use generative AI safely are generated based on a prompt. For example, a prompt might be in the format of "Based on the following information, please generate guidelines for companies to use generative AI safely: While there are some potential security risks associated with using ChatGPT, it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews."

[1001] 6. Display means

[1002] Finally, the generated answers and guidelines are returned to the user's input means and presented to the user via a smartphone application or web browser, allowing the user to review the generated information and take any necessary actions.

[1003] For example, if a user types a question like "I want to use ChatGPT in my company, but are there any security issues?", the system will follow these steps:

[1004] 1. A question is received and the keywords "ChatGPT" and "security" are extracted using natural language processing.

[1005] 2. The database is searched based on the extracted keywords to obtain relevant security policies.

[1006] 3. Specific guidelines are generated using natural language generation tools and generative AI models.

[1007] 4. Users will be shown guidelines such as, "Companies can safely use ChatGPT by meeting the following conditions: 1. Restrict access outside the company network, and 2. We recommend conducting regular security reviews."

[1008] In this way, the system of the present invention provides specific guidelines to enable companies to use generative AI with confidence, reducing security and legal risks.

[1009] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1010] Step 1:

[1011] A user accesses the system via a smartphone application or web browser and enters a question. The question entered is something like, "I want to use ChatGPT in my company, but are there any security issues?" Input: User question. Output: User query data.

[1012] Step 2:

[1013] The server receives the user query data and analyzes the query using natural language processing means. It uses a natural language processing library such as SpaCy to extract important keywords and contextual information from the query. Input: User query data. Output: Extracted keywords and contextual information.

[1014] Step 3:

[1015] The server searches for relevant information through a database connection based on the extracted keywords and context information. This involves sending queries to a database containing corporate rules and legal information. Specifically, it retrieves relevant information from an SQLite database. Input: Extracted keywords and context information. Output: Database search results.

[1016] Step 4:

[1017] The server uses natural language generation to generate user-friendly answers based on information retrieved from the database. A generative AI model such as OpenAI GPT-3 handles this process, providing specialized information in a format that is easy for users to understand. Input: Database search results. Output: Generated natural language answers.

[1018] Step 5:

[1019] The server uses a generative AI model to generate specific guidelines to further develop the generated answers. The generative AI model uses the OpenAI API, and specific guidelines are generated based on the prompt text. For example, the prompt text could be, "Based on the following information, please generate guidelines for the safe use of generative AI by the company: There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews." Input: Generated answer and prompt text. Output: Specific guidelines.

[1020] Step 6:

[1021] The server uses a display means to present the generated guidelines to the user. The generated information is displayed to the user through a smartphone application or a web browser. Input: Generated guidelines. Output: Guidelines displayed to the user.

[1022] example:

[1023] The user enters a question through the application: "I would like to use ChatGPT in my company, but are there any security issues?"

[1024] The server receives and analyzes the question, extracting the keywords "ChatGPT" and "security."

[1025] The server retrieves relevant laws and guidelines from a database based on the extracted keywords.

[1026] The server generates a response based on the acquired information using natural language generation means.

[1027] Using the answers generated by the server, a generative AI model is used to generate specific guidelines.

[1028] The server displays the generated guidelines to the user on the application.

[1029] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1030] The security concierge system incorporating the emotion engine of the present invention is a system for recognizing a user's emotion and providing a more appropriate response based on the emotion. This system includes, as its main components, a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, and an emotion engine.

[1031] System configuration

[1032] 1. User Input Method

[1033] Users access the web interface via a terminal.

[1034] This interface includes a question form and a submit button, allowing users to enter and submit questions about the generated AI.

[1035] 2. Emotion Engine

[1036] The server receives the question sent from the user input means and passes the question to the emotion engine at the same time.

[1037] The emotion engine analyzes the user's emotions from the question's sentence structure and keywords, and determines emotions such as "anxiety," "interest," and "anger."

[1038] The analyzed emotional information is reflected in the subsequent natural language processing means and natural language generation means.

[1039] 3. Natural Language Processing Methods

[1040] The server analyzes the user's question using a natural language processing module along with emotional information from the emotion engine.

[1041] This module extracts important keywords and contextual information from questions, and also takes into account the output of the emotion engine.

[1042] 4. Database connection method

[1043] The server generates and sends a query to the corresponding database based on the keywords and emotional information extracted by natural language processing.

[1044] The database contains corporate security rules and legal information and searches for and returns the appropriate information.

[1045] 5. Natural language generation means

[1046] The server generates a response in a format that is easy for the user to understand based on the information acquired from the database connection means as well as the emotion information output by the emotion engine.

[1047] A natural language generation module takes over this process, making it possible to generate answers that better reflect the user's feelings.

[1048] 6. Display means

[1049] The server formats the generated response in HTML format or similar and sends it to the user's terminal.

[1050] The terminal displays the received answers on a web interface for easy confirmation by the user.

[1051] Specific examples

[1052] 1. The user accesses the Security Concierge web interface on their device.

[1053] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[1054] 3. When the server receives the question, the emotion engine simultaneously analyzes the question and determines the user's emotion as "anxiety."

[1055] 4. The server uses a natural language processing module to extract the important keywords "ChatGPT" and "security."

[1056] 5. The server sends a query to the database based on the keywords and emotion information to search for relevant company rules and legal information.

[1057] 6. The database returns the search results to the server.

[1058] 7. The server uses a natural language generation module to generate answers that also reflect emotional information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the conditions for not needing to be concerned are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[1059] 8. The server sends the generated answer to the user's device.

[1060] 9. The device displays the received answer, and the user can decide on the next action based on the answer (e.g., setting access restrictions outside the company network).

[1061] In this way, a security concierge system combined with an emotion engine can take into account the user's emotions and provide more appropriate and personalized answers.

[1062] The processing flow will be explained below.

[1063] Step 1:

[1064] The user opens a web browser on the device and accesses the Security Concierge web interface.

[1065] Step 2:

[1066] Users enter questions about the generative AI service into a question form on the interface (e.g., "I would like to use ChatGPT in my company, but are there any security issues?").

[1067] Step 3:

[1068] After the user enters the question, he or she clicks the submit button.

[1069] Step 4:

[1070] The terminal sends the user's input to the server as an HTTP request.

[1071] Step 5:

[1072] The server receives the HTTP request and extracts the user's question.

[1073] Step 6:

[1074] The server passes the user's question to the emotion engine, which analyzes the user's emotions.

[1075] Step 7:

[1076] The emotion engine (a component within the server) analyzes the input query and determines the user's emotion from the content (e.g., "anxiety," "interest," "anger," etc.).

[1077] Step 8:

[1078] The server passes the question data to a natural language processing (NLP) module along with the emotion information determined by the emotion engine.

[1079] Step 9:

[1080] The natural language processing (NLP) module analyzes the question and extracts keywords and contextual information (e.g., "ChatGPT," "security").

[1081] Step 10:

[1082] The server generates a search query based on the extracted keywords and emotion information.

[1083] Step 11:

[1084] The server sends the generated search query to the database.

[1085] Step 12:

[1086] The database searches for relevant information (e.g., company rules and legal information) based on the received query.

[1087] Step 13:

[1088] The database returns the search results to the server.

[1089] Step 14:

[1090] The server passes the acquired information and the emotional information from the emotion engine to the natural language generation (NLG) module and requests it to generate an answer.

[1091] Step 15:

[1092] The natural language generation (NLG) module generates a user-friendly answer based on the search results and sentiment information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the conditions for not needing to be concerned are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[1093] Step 16:

[1094] The server converts the generated response into HTML format and sends it to the user's terminal.

[1095] Step 17:

[1096] The terminal displays the received HTML data in a web browser so that the user can easily check it.

[1097] Step 18:

[1098] The user reviews the displayed answers and decides on the next action based on them (e.g., setting restrictions on access outside the company network).

[1099] Example 2

[1100] Next, a description will be given of 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."

[1101] In conventional information search systems, the main purpose is to provide appropriate information in response to user queries. However, because they simply perform keyword matching without considering the user's emotions, it is difficult to provide answers that are in tune with the user's emotions. If a user asks a question harboring a specific emotion, such as anxiety or curiosity, an answer that ignores that emotion is likely to reduce satisfaction. Therefore, the present invention aims to provide a means for analyzing a user's emotions and providing a more appropriate answer based on that emotion.

[1102] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a user input means, a natural language processing means and an emotion analysis means for analyzing the query and emotion, a database connection means for searching for related information based on the analyzed query and emotion, a natural language generation means for generating an easy-to-understand answer based on the acquired information and the user's emotion, and a display means for presenting the generated answer to the user. This makes it possible to analyze the user's emotion and provide an appropriate and personalized answer based on the analysis.

[1103] "User input means" means an interface through which a user accesses the system to input and transmit information.

[1104] "Natural language processing means" refers to technology for analyzing queries entered by users and extracting important keywords and contextual information.

[1105] "Emotion analysis means" refers to a technology for analyzing emotions based on user input and determining a specific emotion (e.g., anxiety, interest, anger, etc.).

[1106] The "database connection means" means a means for transmitting a query to a database to search for related data and obtain information based on the information analyzed by the natural language processing means and the sentiment analysis means.

[1107] "Natural language generation means" refers to a technology that reflects information obtained from a database and user emotional information, and generates answers in an easy-to-understand format based on that information.

[1108] "Display means" refers to an interface for presenting the generated answers to the user, typically a web interface or a screen display.

[1109] A specific example of the invention is described below. This system is a security concierge system that analyzes a user's emotions and provides appropriate answers based on the analysis. This system is mainly composed of a user input means, a natural language processing means, an emotion analysis means, a database connection means, a natural language generation means, and a display means.

[1110] Hardware and software used

[1111] 1. Hardware

[1112] Server: A computer that processes user requests and performs the necessary calculations.

[1113] Terminal: The device from which the user accesses the site (PC, tablet, smartphone, etc.).

[1114] 2. Software

[1115] Web Interface: A UI for users to enter questions.

[1116] Natural language processing module: Software that analyzes the user's question.

[1117] Emotion engine: Software that analyzes user emotions.

[1118] Database: Data storage where company security regulations and legal information are kept.

[1119] Natural Language Generation Module: Software that generates responses.

[1120] Specific operation of the system

[1121] 1. User Input

[1122] The user accesses the system's web interface using a terminal, enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the inquiry form, and presses the submit button.

[1123] 2. User Question and Sentiment Analysis

[1124] The server receives the question data and passes it to the emotion engine, which analyzes the question and determines the emotion, such as "anxiety."

[1125] 3. Natural Language Processing

[1126] The server passes the question, including the emotion information from the emotion engine, to the natural language processing module, which extracts important keywords such as "ChatGPT" and "security" from the question.

[1127] 4. Database query generation and submission

[1128] The server generates a database query based on the extracted keywords and emotion information and sends it to the database.

[1129] 5. Information acquisition

[1130] The database searches for relevant information and returns the results to the server, including information such as terms and security guidelines for using ChatGPT safely within your company.

[1131] 6. Answer generation

[1132] The server uses a natural language generation module to generate a response to the user based on the information retrieved from the database and the output of the emotion engine. For example, it could generate a specific response such as, "There are some security risks associated with using ChatGPT, but it can be used safely if the following conditions are met: 1. Access outside the company network is restricted, and 2. Regular security reviews are required."

[1133] 7. View Answers

[1134] The server formats the generated response in HTML format and sends it to the user's device. The device displays the received response on a web interface, and the user decides what to do next based on the response.

[1135] Specific examples

[1136] A specific example is shown below.

[1137] 1. User Input Example

[1138] "We want to use ChatGPT in our company, but are there any security issues?"

[1139] 2. Example answers generated

[1140] "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met to eliminate them. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews must be conducted."

[1141] The above is an embodiment of the security concierge system incorporating an emotion engine.

[1142] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1143] Step 1: User enters and submits question

[1144] Input: The user accesses the web interface using a terminal, enters the question "I would like to use ChatGPT in my company, but are there any security issues?" and presses the send button.

[1145] Processing: The terminal sends the entered question data to the server.

[1146] Output: The question data is sent to the server.

[1147] Step 2: The server receives the question and analyzes the sentiment

[1148] Input: The server receives the query data sent from the terminal.

[1149] Processing: The server passes the question data to the emotion engine, which analyzes the user's emotions. The emotion engine performs natural language analysis based on the text information and determines emotions such as "anxiety."

[1150] Output: Emotional information (e.g., "anxiety") is generated as the analysis result.

[1151] Step 3: The server passes the question to the natural language processing module for analysis.

[1152] Input: The server passes the question data including the emotion information obtained from the emotion engine to the natural language processing module.

[1153] Processing: The natural language processing module analyzes the question data and extracts important keywords and contextual information (e.g., "ChatGPT" and "security").

[1154] Output: Extracted keywords and context information.

[1155] Step 4: Server generates database query

[1156] Input: The server generates a query to the database based on keywords and sentiment information obtained from the natural language processing module.

[1157] Processing: Generate queries in a specified format (e.g., "ChatGPT usage security risk") based on keywords (e.g., "ChatGPT" and "security") and sentiment information (e.g., "anxiety").

[1158] Output: The generated database query.

[1159] Step 5: Server sends query to database

[1160] Input: The server receives the generated database query.

[1161] Processing: The server sends a database query to the database through the database connection means.

[1162] Output: The database query is passed to the database.

[1163] Step 6: The database searches for the corresponding information and returns it to the server

[1164] Input: The database retrieves information based on the received query.

[1165] Processing: The database searches for information that matches the query and returns the results to the server, such as terms and security guidelines for using ChatGPT safely within your company.

[1166] Output: Search result information.

[1167] Step 7: Server Generates Answer

[1168] Input: The server receives the information retrieved from the database and the emotion information.

[1169] Processing: The server uses a natural language generation module to generate a specific response to the user based on the acquired information and sentiment information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met that eliminate the need for concern. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[1170] Output: The generated answer.

[1171] Step 8: Present the server-generated answer to the user

[1172] Input: The server receives the generated answer.

[1173] Processing: The server formats the response in HTML format or similar and sends it to the user's terminal.

[1174] Output: The formatted response data is sent to the terminal.

[1175] Step 9: Your device will display your answer

[1176] Input: The terminal receives the response data sent from the server.

[1177] Processing: The terminal displays the received response on the web interface.

[1178] Output: The user can check the answer on the display screen.

[1179] (Application example 2)

[1180] Next, a description will be given of Application 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."

[1181] Conventional security concierge systems can only generate simple answers to user input, making it difficult to provide personalized answers that take the user's emotions into account. This has led to the problem that they are unable to provide a highly satisfying service, especially to users who have concerns or questions.

[1182] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, and an emotion engine. This makes it possible to provide a personalized answer that takes into account the user's emotions.

[1183] The "user input means" is an interface for a user to input a query.

[1184] "Natural language processing means" is a means for analyzing a user's query and extracting keywords and contextual information.

[1185] The "database connection means" is a means for searching and acquiring related information from a database based on keywords extracted by the natural language processing means.

[1186] The "natural language generation means" is a means for generating a response in a format that is easy for the user to understand, based on information acquired from the database connection means.

[1187] The "display means" is a device or system for presenting the generated answer to the user.

[1188] An "emotion engine" is a system that recognizes emotions contained in user input and provides that emotion information to other processing means.

[1189] The present invention provides a security concierge system that provides personalized answers taking into account the emotions of a user. The system includes a user input unit, an emotion engine, a natural language processing unit, a database connection unit, a natural language generation unit, and a display unit.

[1190] System configuration

[1191] 1. User Input Method

[1192] Users enter queries through a smartphone application or web interface, which also includes auto-completion and voice input.

[1193] 2. Emotion Engine

[1194] The server passes queries received from the user's input means to the emotion engine, which analyzes the user's input text and voice data and classifies the user's emotions into categories such as "anxiety," "anger," and "interest." This analysis uses emotion analysis APIs such as IBM Watson's Tone Analyzer.

[1195] 3. Natural Language Processing Methods

[1196] The server analyzes the user's query using a natural language processing module along with sentiment information from the sentiment engine, using the Google Natural Language API to extract key keywords and contextual information.

[1197] 4. Database connection method

[1198] Based on the keywords and sentiment information extracted through natural language processing, the server queries an internal security database, which contains information on the company's security regulations and laws, to retrieve relevant information.

[1199] 5. Natural language generation means

[1200] The server generates answers in a user-friendly format using OpenAI's GPT-3 or similar software based on the information obtained from the database connection means and the output of the emotion engine. Generation based on emotion information enables more personalized answers.

[1201] 6. Display means

[1202] The server formats the generated answer in HTML format or similar and sends it to the user's device, which then displays the received answer on a web interface or application UI.

[1203] Specific examples

[1204] If a user enters the query, "I want to use ChatGPT in my company, but are there any security issues?", the system will behave as follows:

[1205] 1. Upon receiving the query, the server sends a question to the emotion engine and determines the user's emotion as "anxiety."

[1206] 2. The server uses a natural language processing module to extract important keywords such as "ChatGPT" and "security."

[1207] 3. The server sends a query to the database based on the extracted keywords and emotion information to obtain relevant company rules and legal information.

[1208] 4. The server generates a response using OpenAI's GPT-3 based on the acquired information and emotion information. The response generated is, "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required."

[1209] 5. The server sends the generated answer to the user's smartphone and displays it on the application.

[1210] Prompt Sentence Examples

[1211] The user is feeling anxious. Please provide gentle advice based on the following information:

[1212] 1. General risks in implementing new security measures

[1213] 2. Regular security reviews are necessary

[1214] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1215] Step 1:

[1216] The user opens the application on their smartphone and inputs a question. The input question is sent to the device as text or voice. The input data includes specific questions such as, "We want to use ChatGPT in our company, but are there any security issues?" This data is sent to the server, which initiates the next processing step.

[1217] Step 2:

[1218] The server passes the query received from the user's input means to the emotion engine. The emotion engine analyzes the text data of the received query and classifies the user's emotion. Specifically, it uses IBM Watson's Tone Analyzer to determine emotions such as "anxiety," "anger," and "interest" from the input question. Emotional information such as "anxiety" is output as the analysis result.

[1219] Step 3:

[1220] The server then sends the query, along with the emotion information from the emotion engine, to the natural language processing module, which uses the Google Natural Language API to analyze the query and extract important keywords and contextual information. For example, keywords like "ChatGPT" and "security" are extracted and sent to the next processing step.

[1221] Step 4:

[1222] The server uses the extracted keywords and emotion information to search for relevant information using a database connection method. The database stores information on corporate security rules and laws. An SQL query is sent to this database to retrieve the appropriate information. For example, information on "security risks regarding the internal use of ChatGPT" is returned as a search result.

[1223] Step 5:

[1224] The server uses natural language generation to generate responses to users based on the information and emotional information obtained from the database connection means. It uses OpenAI's GPT-3 model to create prompts based on the input data and generate emotional responses. A specific example of output might be something like, "There are some risks associated with using ChatGPT, but it can be used if appropriate measures are taken."

[1225] Step 6:

[1226] The server formats the generated answer in HTML format and sends it to the user's device. The device displays the received answer on the application. The user can then decide what to do next based on the answer.

[1227] Specific actions

[1228] Users use their smartphones to enter and submit questions.

[1229] The server uses an emotion engine to analyze the user's emotions.

[1230] The server uses a natural language processing module to extract important keywords.

[1231] The server queries the database to retrieve the appropriate information.

[1232] The server uses natural language generation means to generate answers that are easy for the user to understand.

[1233] The server sends the generated response to the terminal, which displays it.

[1234] 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 audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio 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 audio data.

[1235] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with 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.

[1236] 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.

[1237] [Fourth embodiment]

[1238] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1239] 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.

[1240] 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 WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1241] 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. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1242] 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 audio in accordance with instructions from the processor 46.

[1243] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and 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 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1244] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1245] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. 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.

[1246] Fig. 8 shows an example of the 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.

[1247] 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.

[1248] 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.

[1249] In the robot 414, the processor 46 performs the reception output process. 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.

[1250] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. 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."

[1251] The security concierge system of the present invention is a system for reducing security and legal risks associated with the use of generative AI in companies, allowing companies to use generative AI with peace of mind. This system includes, as its main components, a user input means, a natural language processing means, a database connection means, a natural language generation means, and a display means.

[1252] System configuration

[1253] 1. User Input Method

[1254] Users access the web interface via a terminal.

[1255] This interface includes a question form and a submit button, allowing users to enter and submit questions about the generated AI.

[1256] 2. Natural Language Processing Methods

[1257] The server receives a question sent from a user input means.

[1258] The received question is analyzed by a natural language processing module to extract the most important keywords and context.

[1259] 3. Database connection method

[1260] The server uses the keywords extracted by natural language processing to send queries to a database to search for relevant information.

[1261] This database stores information about corporate security rules and regulations.

[1262] 4. Natural language generation means

[1263] The server generates an answer that is easy for the user to understand based on the information retrieved from the database.

[1264] A natural language generation module takes over this process to generate the most appropriate answer to the query.

[1265] 5. Display means

[1266] The server sends the generated answer to the user's terminal.

[1267] The terminal displays the received answers on a web interface for easy confirmation by the user.

[1268] Specific examples

[1269] 1. The user accesses the Security Concierge web interface on their device.

[1270] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[1271] 3. The server receives the question, and the natural language processing module extracts the key keywords "ChatGPT" and "security."

[1272] 4. Based on these keywords, the server searches the database for relevant company rules and legal information.

[1273] 5. The database returns the search results (e.g., the company's security policy regarding the use of ChatGPT).

[1274] 6. The server uses a natural language generation module to generate a response that states, "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, and 2. Regular security reviews are conducted."

[1275] 7. The server formats this response in HTML or other format and sends it to the terminal.

[1276] 8. The terminal displays the received answer to the user, who can then take necessary action based on the answer.

[1277] In this way, the system can quickly and accurately answer user questions about the use of generative AI and provide guidelines for companies to use generative AI with confidence.

[1278] The processing flow will be explained below.

[1279] Step 1:

[1280] The user opens a web browser on the device and accesses the Security Concierge web interface.

[1281] Step 2:

[1282] Users enter questions about the generative AI service into a question form on the interface (e.g., "I would like to use ChatGPT in my company, but are there any security issues?").

[1283] Step 3:

[1284] After the user enters the question, he or she clicks the submit button.

[1285] Step 4:

[1286] The terminal sends the user's input to the server as an HTTP request.

[1287] Step 5:

[1288] The server receives the HTTP request and extracts the user's question.

[1289] Step 6:

[1290] The server passes the question to a natural language processing (NLP) module to begin parsing it.

[1291] Step 7:

[1292] The natural language processing (NLP) module analyzes the question and extracts keywords and contextual information (e.g., "ChatGPT," "security").

[1293] Step 8:

[1294] The server generates a search query based on the extracted keywords.

[1295] Step 9:

[1296] The server sends the generated search query to the database.

[1297] Step 10:

[1298] The database searches for relevant company rules and legal information based on the received query.

[1299] Step 11:

[1300] The database returns the search results to the server.

[1301] Step 12:

[1302] The server passes the retrieved information to a natural language generation (NLG) module to begin generating an answer.

[1303] Step 13:

[1304] The Natural Language Generation (NLG) module generates answers based on the search results in a user-friendly format (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted").

[1305] Step 14:

[1306] The server converts the generated response into HTML format and sends it to the user's terminal.

[1307] Step 15:

[1308] The terminal displays the received HTML data in a web browser.

[1309] Step 16:

[1310] The user confirms the answers on the screen and decides on the next action (e.g., setting restrictions on access outside the company network).

[1311] Example 1

[1312] Next, a description will be given of 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."

[1313] The use of generative AI in companies involves security and legal risks, making it difficult to use safely. This can delay the introduction of generative AI or cause problems after its introduction. The objective of this invention is to provide a system for reducing these risks and operating generative AI safely and efficiently.

[1314] 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.

[1315] In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, a means for the user input means to receive a user input via a web interface and send a query by operating a send button, a means for the natural language processing means to analyze keywords and context information in the query using a natural language processing module, a means for the natural language generation means to generate an answer based on information obtained from a database using the natural language generation module, and a means for the display means to receive the generated answer and display it on the web interface. This reduces security and legal risks when using generative AI, allowing companies to use generative AI with peace of mind.

[1316] "User input means" refers to a means by which a user provides input to the system.

[1317] "Natural language processing means" is a means for analyzing queries received from users and extracting important keywords and contextual information.

[1318] The "database connection means" is a means for accessing a database based on a query and searching for related information.

[1319] "Natural language generation means" is a means for generating appropriate answers to users based on acquired information.

[1320] The "display means" is a means for presenting the generated answer to the user.

[1321] A "web interface" is an interface through which a user interacts with a system via a web browser.

[1322] A "natural language processing module" is a software module for analyzing input text and extracting keywords and contextual information.

[1323] A "natural language generation module" is a software module for generating answers in natural language based on analyzed information.

[1324] A "server" is a computer that performs the central processing of the system, and is a device that receives input from users and analyzes, processes, and generates responses.

[1325] A "query" refers to a question or request made by a user to a system.

[1326] The security concierge system of the present invention is intended to enable the safe use of generative AI in companies and to reduce security and legal risks. This system is implemented using the following hardware and software.

[1327] User Input Method

[1328] A user accesses the web interface of the security concierge system using a web browser. This interface includes a question form and a submit button that serves to transmit the query entered by the user to the server.

[1329] Natural language processing tools

[1330] The server receives queries sent by users. This processing is done using web server software such as Nginx or Apache. The received queries are processed using web frameworks implemented in Python, such as Django or Flask. The server then calls a natural language processing module to analyze the queries. This module uses natural language processing libraries such as NLTK or spaCy.

[1331] Database connection method

[1332] The server uses the keywords and contextual information extracted from the query to send an SQL query to a database. This database stores information about corporate security rules and regulations, and is typically an RDBMS such as MySQL or PostgreSQL. The server accesses the database using a database connection library (e.g., SQLAlchemy or Psycopg2) to retrieve the relevant information.

[1333] Natural language generation means

[1334] The server uses a natural language generation module based on the information retrieved from the database to generate answers for the user. This module uses OpenAI's GPT-3 or GPT-4, for example. Answers are generated in a format that is easy for the user to understand.

[1335] Display means

[1336] The server formats the generated response in HTML format or similar and sends it to the user's device, which then displays the response on a web browser so that the user can easily view it.

[1337] Specific example explanation

[1338] A specific example of the operation of the system is shown below.

[1339] 1. The user accesses the Security Concierge web interface using a terminal.

[1340] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[1341] 3. The server receives this question, and the natural language processing module extracts the keywords "ChatGPT" and "security."

[1342] 4. The server queries the database based on these keywords to find relevant information.

[1343] 5. The database returns the search results (e.g., the company's security policy regarding the use of ChatGPT).

[1344] 6. The server uses a natural language generation module to generate a response that states, "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted."

[1345] 7. The server formats this response in HTML format and sends it to the terminal.

[1346] 8. The device displays this response in the display area on the web browser so that the user can confirm it.

[1347] In this way, the system can provide quick and accurate answers to user questions about the use of generative AI and provide guidelines for companies to use generative AI with confidence.

[1348] Prompt Sentence Examples

[1349] Here are some examples of prompts to input to a generative AI model:

[1350] "What are the security risks when using ChatGPT in-house?"

[1351] "I want to know about the legal risks and countermeasures when introducing generative AI in-house."

[1352] "What are the best practices for enterprises using generative AI?"

[1353] Using this prompt, the system can generate an appropriate response to help the user obtain the information they need.

[1354] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1355] Step 1:

[1356] A user uses a terminal to access the web interface of the security concierge system via a web browser. This is equivalent to issuing an HTTP request and accessing the server. The user enters a question in the question form and clicks the send button. This operation causes the user input means to send the question data to the server.

[1357] Input: A query entered by a user into an input form (e.g., "I want to use ChatGPT in my company, but are there any security issues?")

[1358] Output: Query data sent from the user input means to the server

[1359] Step 2:

[1360] The server receives queries sent by users via web server software (e.g., Nginx or Apache) and a framework implemented in Python (e.g., Django or Flask) handles the processing of the queries. The server then passes the queries to a natural language processor.

[1361] Input: Query data sent from user input means

[1362] Output: Unparsed query data passed to natural language processing tools

[1363] Step 3:

[1364] The server analyzes the query using natural language processing. It uses a natural language processing module (e.g., NLTK or spaCy) to extract important keywords and contextual information from the query. Specifically, it performs processes such as tokenization, morphological analysis, and tagging. For example, it extracts the keywords "ChatGPT" and "security" from the query, "We want to use ChatGPT in our company, but are there any security issues?"

[1365] Input: Unparsed query data

[1366] Output: Extracted keywords and context information (e.g., "ChatGPT", "security")

[1367] Step 4:

[1368] The server searches for relevant information using a database connection method based on the extracted keywords. Specifically, it generates an SQL query and accesses the database (MySQL or PostgreSQL) using a database connection library (e.g., SQLAlchemy or Psycopg2). An example of the query content is "SELECT FROM security_policies WHERE keywords LIKE '%ChatGPT%' AND keywords LIKE '%security%'".

[1369] Input: Extracted keywords and context information

[1370] Output: Relevant information retrieved from the database (e.g., security policy regarding ChatGPT usage)

[1371] Step 5:

[1372] The server uses natural language generation to generate answers based on information retrieved from the database. It uses natural language generation modules (such as OpenAI's GPT-3 or GPT-4) to format the retrieved information in an easy-to-understand format.

[1373] Input: Relevant information retrieved from the database

[1374] Output: Answer generated by natural language generation (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are conducted.")

[1375] Step 6:

[1376] The server formats the generated response into an appropriate HTML format and sends it to the user's device as an HTTP response, setting the response headers and status code appropriately.

[1377] Input: Answer generated by a natural language generator

[1378] Output: HTML formatted response data sent to the terminal as an HTTP response

[1379] Step 7:

[1380] The terminal displays the received HTML response in a web browser using an HTML rendering engine, and the content is displayed in a format that the user can visually confirm.

[1381] Input: HTML formatted response data sent from the server

[1382] Output: Answer displayed in web browser (e.g., "ChatGPT may pose some security risks, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews are performed.")

[1383] As described above, the overall processing of the system is realized in a series of steps, and at each step the input data is appropriately processed and analyzed, ultimately providing the appropriate answer to the user.

[1384] (Application example 1)

[1385] Next, a description will be given of Application 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."

[1386] There is a need to manage the security and legal risks associated with the use of generative AI and provide an environment in which companies can use it with peace of mind. However, current systems often lack specific guidelines for specific risks associated with the use of generative AI and information for legal compliance. Furthermore, there are no systems in place that can assess risks and generate guidelines in a format that is easy for users to understand. This has resulted in a lack of comprehensive solutions for reducing the risks that arise when companies operate generative AI.

[1387] 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.

[1388] In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a means for generating guidelines using a generative AI model, and a display means, which makes it possible to search for information on relevant laws and guidelines based on a question from a user, generate a risk assessment and specific guidelines using the generative AI model, and display them in an easy-to-understand manner for the user.

[1389] "User input means" refers to an interface for accepting questions or requests from users, and includes smartphone applications, web browsers, and other input devices.

[1390] "Natural language processing means" is a technology for analyzing queries (questions or requests) obtained from user input means and extracting important keywords and contextual information.

[1391] The "database connection means" has a function of accessing a database to search for related information based on a query analyzed by the natural language processing means.

[1392] The "natural language generation means" is a technology for generating answers that are easy for the user to understand based on information acquired by the database connection means.

[1393] The "display means" is an interface for presenting the answer generated by the natural language generation means to the user, and includes a smartphone screen, a computer display, and other display devices.

[1394] "Laws and guidelines" includes rules, regulations, and guidelines that companies must comply with when using generative AI.

[1395] A "generative AI model" is an artificial intelligence technology that generates appropriate answers and guidelines based on information from users.

[1396] A "prompt" is a sentence or instruction that serves as a starting point for a generative AI model to generate appropriate answers or guidelines.

[1397] This invention provides a system for implementing security risk assessment and guideline generation. This system is composed of a user input means, a natural language processing means, a database connection means, a natural language generation means, a guideline generation means using a generative AI model, and a display means. Each component and its operation procedure will be specifically explained below.

[1398] 1. User Input Method

[1399] The user input means provides an interface for users to input questions or requests to the generative AI. Specifically, this corresponds to a smartphone application, a web browser, etc. When the user inputs a question, the query is sent to the system.

[1400] 2. Natural Language Processing Methods

[1401] The server is equipped with a natural language processing unit to analyze queries received from users, using technologies such as spaCy to extract important keywords and contextual information from the queries.

[1402] 3. Database connection method

[1403] Based on the extracted keywords and context information, the server searches for relevant information using a database connection method. The database stores corporate rules and legal information. For example, an SQLite database is used.

[1404] 4. Natural language generation means

[1405] The information retrieved from the database is used by a natural language generation system on the server to generate answers in a format that is easy for users to understand. This process uses generative AI models such as OpenAI GPT-3, which allows information to be provided in a format that is easy for users to understand, even if the content is technical.

[1406] 5. Guideline generation method using generative AI model

[1407] Furthermore, it includes a means for generating specific guidelines using a generative AI model based on the generated information. The generative AI model uses the OpenAI API, and specific guidelines for companies to use generative AI safely are generated based on a prompt. For example, a prompt might be in the format of "Based on the following information, please generate guidelines for companies to use generative AI safely: While there are some potential security risks associated with using ChatGPT, it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews."

[1408] 6. Display means

[1409] Finally, the generated answers and guidelines are returned to the user's input means and presented to the user via a smartphone application or web browser, allowing the user to review the generated information and take any necessary actions.

[1410] For example, if a user types a question like "I want to use ChatGPT in my company, but are there any security issues?", the system will follow these steps:

[1411] 1. A question is received and the keywords "ChatGPT" and "security" are extracted using natural language processing.

[1412] 2. The database is searched based on the extracted keywords to obtain relevant security policies.

[1413] 3. Specific guidelines are generated using natural language generation tools and generative AI models.

[1414] 4. Users will be shown guidelines such as, "Companies can safely use ChatGPT by meeting the following conditions: 1. Restrict access outside the company network, and 2. We recommend conducting regular security reviews."

[1415] In this way, the system of the present invention provides specific guidelines to enable companies to use generative AI with confidence, reducing security and legal risks.

[1416] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1417] Step 1:

[1418] A user accesses the system via a smartphone application or web browser and enters a question. The question entered is something like, "I want to use ChatGPT in my company, but are there any security issues?" Input: User question. Output: User query data.

[1419] Step 2:

[1420] The server receives the user query data and analyzes the query using natural language processing means. It uses a natural language processing library such as SpaCy to extract important keywords and contextual information from the query. Input: User query data. Output: Extracted keywords and contextual information.

[1421] Step 3:

[1422] The server searches for relevant information through a database connection based on the extracted keywords and context information. This involves sending queries to a database containing corporate rules and legal information. Specifically, it retrieves relevant information from an SQLite database. Input: Extracted keywords and context information. Output: Database search results.

[1423] Step 4:

[1424] The server uses natural language generation to generate user-friendly answers based on information retrieved from the database. A generative AI model such as OpenAI GPT-3 handles this process, providing specialized information in a format that is easy for users to understand. Input: Database search results. Output: Generated natural language answers.

[1425] Step 5:

[1426] The server uses a generative AI model to generate specific guidelines to further develop the generated answers. The generative AI model uses the OpenAI API, and specific guidelines are generated based on the prompt text. For example, the prompt text could be, "Based on the following information, please generate guidelines for the safe use of generative AI by the company: There are some security risks associated with using ChatGPT, but it can be used if the following conditions are met: 1. Restrict access outside the company network, 2. Regular security reviews." Input: Generated answer and prompt text. Output: Specific guidelines.

[1427] Step 6:

[1428] The server uses a display means to present the generated guidelines to the user. The generated information is displayed to the user through a smartphone application or a web browser. Input: Generated guidelines. Output: Guidelines displayed to the user.

[1429] example:

[1430] The user enters a question through the application: "I would like to use ChatGPT in my company, but are there any security issues?"

[1431] The server receives and analyzes the question, extracting the keywords "ChatGPT" and "security."

[1432] The server retrieves relevant laws and guidelines from a database based on the extracted keywords.

[1433] The server generates a response based on the acquired information using natural language generation means.

[1434] Using the answers generated by the server, a generative AI model is used to generate specific guidelines.

[1435] The server displays the generated guidelines to the user on the application.

[1436] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1437] The security concierge system incorporating the emotion engine of the present invention is a system for recognizing a user's emotion and providing a more appropriate response based on the emotion. This system includes, as its main components, a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, and an emotion engine.

[1438] System configuration

[1439] 1. User Input Method

[1440] Users access the web interface via a terminal.

[1441] This interface includes a question form and a submit button, allowing users to enter and submit questions about the generated AI.

[1442] 2. Emotion Engine

[1443] The server receives the question sent from the user input means and passes the question to the emotion engine at the same time.

[1444] The emotion engine analyzes the user's emotions from the question's sentence structure and keywords, and determines emotions such as "anxiety," "interest," and "anger."

[1445] The analyzed emotional information is reflected in the subsequent natural language processing means and natural language generation means.

[1446] 3. Natural Language Processing Methods

[1447] The server analyzes the user's question using a natural language processing module along with emotional information from the emotion engine.

[1448] This module extracts important keywords and contextual information from questions, and also takes into account the output of the emotion engine.

[1449] 4. Database connection method

[1450] The server generates and sends a query to the corresponding database based on the keywords and emotional information extracted by natural language processing.

[1451] The database contains corporate security rules and legal information and searches for and returns the appropriate information.

[1452] 5. Natural language generation means

[1453] The server generates a response in a format that is easy for the user to understand based on the information acquired from the database connection means as well as the emotion information output by the emotion engine.

[1454] A natural language generation module takes over this process, making it possible to generate answers that better reflect the user's feelings.

[1455] 6. Display means

[1456] The server formats the generated response in HTML format or similar and sends it to the user's terminal.

[1457] The terminal displays the received answers on a web interface for easy confirmation by the user.

[1458] Specific examples

[1459] 1. The user accesses the Security Concierge web interface on their device.

[1460] 2. The user enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the question form and presses the submit button.

[1461] 3. When the server receives the question, the emotion engine simultaneously analyzes the question and determines the user's emotion as "anxiety."

[1462] 4. The server uses a natural language processing module to extract the important keywords "ChatGPT" and "security."

[1463] 5. The server sends a query to the database based on the keywords and emotion information to search for relevant company rules and legal information.

[1464] 6. The database returns the search results to the server.

[1465] 7. The server uses a natural language generation module to generate answers that also reflect emotional information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the conditions for not needing to be concerned are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[1466] 8. The server sends the generated answer to the user's device.

[1467] 9. The device displays the received answer, and the user can decide on the next action based on the answer (e.g., setting access restrictions outside the company network).

[1468] In this way, a security concierge system combined with an emotion engine can take into account the user's emotions and provide more appropriate and personalized answers.

[1469] The processing flow will be explained below.

[1470] Step 1:

[1471] The user opens a web browser on the device and accesses the Security Concierge web interface.

[1472] Step 2:

[1473] Users enter questions about the generative AI service into a question form on the interface (e.g., "I would like to use ChatGPT in my company, but are there any security issues?").

[1474] Step 3:

[1475] After the user enters the question, he or she clicks the submit button.

[1476] Step 4:

[1477] The terminal sends the user's input to the server as an HTTP request.

[1478] Step 5:

[1479] The server receives the HTTP request and extracts the user's question.

[1480] Step 6:

[1481] The server passes the user's question to the emotion engine, which analyzes the user's emotions.

[1482] Step 7:

[1483] The emotion engine (a component within the server) analyzes the input query and determines the user's emotion from the content (e.g., "anxiety," "interest," "anger," etc.).

[1484] Step 8:

[1485] The server passes the question data to a natural language processing (NLP) module along with the emotion information determined by the emotion engine.

[1486] Step 9:

[1487] The natural language processing (NLP) module analyzes the question and extracts keywords and contextual information (e.g., "ChatGPT," "security").

[1488] Step 10:

[1489] The server generates a search query based on the extracted keywords and emotion information.

[1490] Step 11:

[1491] The server sends the generated search query to the database.

[1492] Step 12:

[1493] The database searches for relevant information (e.g., company rules and legal information) based on the received query.

[1494] Step 13:

[1495] The database returns the search results to the server.

[1496] Step 14:

[1497] The server passes the acquired information and the emotional information from the emotion engine to the natural language generation (NLG) module and requests it to generate an answer.

[1498] Step 15:

[1499] The natural language generation (NLG) module generates a user-friendly answer based on the search results and sentiment information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if the conditions for not needing to be concerned are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[1500] Step 16:

[1501] The server converts the generated response into HTML format and sends it to the user's terminal.

[1502] Step 17:

[1503] The terminal displays the received HTML data in a web browser so that the user can easily check it.

[1504] Step 18:

[1505] The user reviews the displayed answers and decides on the next action based on them (e.g., setting restrictions on access outside the company network).

[1506] Example 2

[1507] Next, a description will be given of 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."

[1508] In conventional information search systems, the main purpose is to provide appropriate information in response to user queries. However, because they simply perform keyword matching without considering the user's emotions, it is difficult to provide answers that are in tune with the user's emotions. If a user asks a question harboring a specific emotion, such as anxiety or curiosity, an answer that ignores that emotion is likely to reduce satisfaction. Therefore, the present invention aims to provide a means for analyzing a user's emotions and providing a more appropriate answer based on that emotion.

[1509] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a user input means, a natural language processing means and an emotion analysis means for analyzing the query and emotion, a database connection means for searching for related information based on the analyzed query and emotion, a natural language generation means for generating an easy-to-understand answer based on the acquired information and the user's emotion, and a display means for presenting the generated answer to the user. This makes it possible to analyze the user's emotion and provide an appropriate and personalized answer based on the analysis.

[1510] "User input means" means an interface through which a user accesses the system to input and transmit information.

[1511] "Natural language processing means" refers to technology for analyzing queries entered by users and extracting important keywords and contextual information.

[1512] "Emotion analysis means" refers to a technology for analyzing emotions based on user input and determining a specific emotion (e.g., anxiety, interest, anger, etc.).

[1513] The "database connection means" means a means for transmitting a query to a database to search for related data and obtain information based on the information analyzed by the natural language processing means and the sentiment analysis means.

[1514] "Natural language generation means" refers to a technology that reflects information obtained from a database and user emotional information, and generates answers in an easy-to-understand format based on that information.

[1515] "Display means" refers to an interface for presenting the generated answers to the user, typically a web interface or a screen display.

[1516] A specific example of the invention is described below. This system is a security concierge system that analyzes a user's emotions and provides appropriate answers based on the analysis. This system is mainly composed of a user input means, a natural language processing means, an emotion analysis means, a database connection means, a natural language generation means, and a display means.

[1517] Hardware and software used

[1518] 1. Hardware

[1519] Server: A computer that processes user requests and performs the necessary calculations.

[1520] Terminal: The device from which the user accesses the site (PC, tablet, smartphone, etc.).

[1521] 2. Software

[1522] Web Interface: A UI for users to enter questions.

[1523] Natural language processing module: Software that analyzes the user's question.

[1524] Emotion engine: Software that analyzes user emotions.

[1525] Database: Data storage where company security regulations and legal information are kept.

[1526] Natural Language Generation Module: Software that generates responses.

[1527] Specific operation of the system

[1528] 1. User Input

[1529] The user accesses the system's web interface using a terminal, enters the question "I would like to use ChatGPT in my company, but are there any security issues?" into the inquiry form, and presses the submit button.

[1530] 2. User Question and Sentiment Analysis

[1531] The server receives the question data and passes it to the emotion engine, which analyzes the question and determines the emotion, such as "anxiety."

[1532] 3. Natural Language Processing

[1533] The server passes the question, including the emotion information from the emotion engine, to the natural language processing module, which extracts important keywords such as "ChatGPT" and "security" from the question.

[1534] 4. Database query generation and submission

[1535] The server generates a database query based on the extracted keywords and emotion information and sends it to the database.

[1536] 5. Information acquisition

[1537] The database searches for relevant information and returns the results to the server, including information such as terms and security guidelines for using ChatGPT safely within your company.

[1538] 6. Answer generation

[1539] The server uses a natural language generation module to generate a response to the user based on the information retrieved from the database and the output of the emotion engine. For example, it could generate a specific response such as, "There are some security risks associated with using ChatGPT, but it can be used safely if the following conditions are met: 1. Access outside the company network is restricted, and 2. Regular security reviews are required."

[1540] 7. View Answers

[1541] The server formats the generated response in HTML format and sends it to the user's device. The device displays the received response on a web interface, and the user decides what to do next based on the response.

[1542] Specific examples

[1543] A specific example is shown below.

[1544] 1. User Input Example

[1545] "We want to use ChatGPT in our company, but are there any security issues?"

[1546] 2. Example answers generated

[1547] "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met to eliminate them. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews must be conducted."

[1548] The above is an embodiment of the security concierge system incorporating an emotion engine.

[1549] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1550] Step 1: User enters and submits question

[1551] Input: The user accesses the web interface using a terminal, enters the question "I would like to use ChatGPT in my company, but are there any security issues?" and presses the send button.

[1552] Processing: The terminal sends the entered question data to the server.

[1553] Output: The question data is sent to the server.

[1554] Step 2: The server receives the question and analyzes the sentiment

[1555] Input: The server receives the query data sent from the terminal.

[1556] Processing: The server passes the question data to the emotion engine, which analyzes the user's emotions. The emotion engine performs natural language analysis based on the text information and determines emotions such as "anxiety."

[1557] Output: Emotional information (e.g., "anxiety") is generated as the analysis result.

[1558] Step 3: The server passes the question to the natural language processing module for analysis.

[1559] Input: The server passes the question data including the emotion information obtained from the emotion engine to the natural language processing module.

[1560] Processing: The natural language processing module analyzes the question data and extracts important keywords and contextual information (e.g., "ChatGPT" and "security").

[1561] Output: Extracted keywords and context information.

[1562] Step 4: Server generates database query

[1563] Input: The server generates a query to the database based on keywords and sentiment information obtained from the natural language processing module.

[1564] Processing: Generate queries in a specified format (e.g., "ChatGPT usage security risk") based on keywords (e.g., "ChatGPT" and "security") and sentiment information (e.g., "anxiety").

[1565] Output: The generated database query.

[1566] Step 5: Server sends query to database

[1567] Input: The server receives the generated database query.

[1568] Processing: The server sends a database query to the database through the database connection means.

[1569] Output: The database query is passed to the database.

[1570] Step 6: The database searches for the corresponding information and returns it to the server

[1571] Input: The database retrieves information based on the received query.

[1572] Processing: The database searches for information that matches the query and returns the results to the server, such as terms and security guidelines for using ChatGPT safely within your company.

[1573] Output: Search result information.

[1574] Step 7: Server Generates Answer

[1575] Input: The server receives the information retrieved from the database and the emotion information.

[1576] Processing: The server uses a natural language generation module to generate a specific response to the user based on the acquired information and sentiment information (e.g., "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met that eliminate the need for concern. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required.").

[1577] Output: The generated answer.

[1578] Step 8: Present the server-generated answer to the user

[1579] Input: The server receives the generated answer.

[1580] Processing: The server formats the response in HTML format or similar and sends it to the user's terminal.

[1581] Output: The formatted response data is sent to the terminal.

[1582] Step 9: Your device will display your answer

[1583] Input: The terminal receives the response data sent from the server.

[1584] Processing: The terminal displays the received response on the web interface.

[1585] Output: The user can check the answer on the display screen.

[1586] (Application example 2)

[1587] Next, a description will be given of Application 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."

[1588] Conventional security concierge systems can only generate simple answers to user input, making it difficult to provide personalized answers that take the user's emotions into account. This has led to the problem that they are unable to provide a highly satisfying service, especially to users who have concerns or questions.

[1589] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a user input means, a natural language processing means, a database connection means, a natural language generation means, a display means, and an emotion engine. This makes it possible to provide a personalized answer that takes into account the user's emotions.

[1590] The "user input means" is an interface for a user to input a query.

[1591] "Natural language processing means" is a means for analyzing a user's query and extracting keywords and contextual information.

[1592] The "database connection means" is a means for searching and acquiring related information from a database based on keywords extracted by the natural language processing means.

[1593] The "natural language generation means" is a means for generating a response in a format that is easy for the user to understand, based on information acquired from the database connection means.

[1594] The "display means" is a device or system for presenting the generated answer to the user.

[1595] An "emotion engine" is a system that recognizes emotions contained in user input and provides that emotion information to other processing means.

[1596] The present invention provides a security concierge system that provides personalized answers taking into account the emotions of a user. The system includes a user input unit, an emotion engine, a natural language processing unit, a database connection unit, a natural language generation unit, and a display unit.

[1597] System configuration

[1598] 1. User Input Method

[1599] Users enter queries through a smartphone application or web interface, which also includes auto-completion and voice input.

[1600] 2. Emotion Engine

[1601] The server passes queries received from the user's input means to the emotion engine, which analyzes the user's input text and voice data and classifies the user's emotions into categories such as "anxiety," "anger," and "interest." This analysis uses emotion analysis APIs such as IBM Watson's Tone Analyzer.

[1602] 3. Natural Language Processing Methods

[1603] The server analyzes the user's query using a natural language processing module along with sentiment information from the sentiment engine, using the Google Natural Language API to extract key keywords and contextual information.

[1604] 4. Database connection method

[1605] Based on the keywords and sentiment information extracted through natural language processing, the server queries an internal security database, which contains information on the company's security regulations and laws, to retrieve relevant information.

[1606] 5. Natural language generation means

[1607] The server generates answers in a user-friendly format using OpenAI's GPT-3 or similar software based on the information obtained from the database connection means and the output of the emotion engine. Generation based on emotion information enables more personalized answers.

[1608] 6. Display means

[1609] The server formats the generated answer in HTML format or similar and sends it to the user's device, which then displays the received answer on a web interface or application UI.

[1610] Specific examples

[1611] If a user enters the query, "I want to use ChatGPT in my company, but are there any security issues?", the system will behave as follows:

[1612] 1. Upon receiving the query, the server sends a question to the emotion engine and determines the user's emotion as "anxiety."

[1613] 2. The server uses a natural language processing module to extract important keywords such as "ChatGPT" and "security."

[1614] 3. The server sends a query to the database based on the extracted keywords and emotion information to obtain relevant company rules and legal information.

[1615] 4. The server generates a response using OpenAI's GPT-3 based on the acquired information and emotion information. The response generated is, "There are some security risks associated with using ChatGPT, but it can be used if certain conditions are met. Specifically, 1. Access outside the company network must be restricted, and 2. Regular security reviews are required."

[1616] 5. The server sends the generated answer to the user's smartphone and displays it on the application.

[1617] Prompt Sentence Examples

[1618] The user is feeling anxious. Please provide gentle advice based on the following information:

[1619] 1. General risks in implementing new security measures

[1620] 2. Regular security reviews are necessary

[1621] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1622] Step 1:

[1623] The user opens the application on their smartphone and inputs a question. The input question is sent to the device as text or voice. The input data includes specific questions such as, "We want to use ChatGPT in our company, but are there any security issues?" This data is sent to the server, which initiates the next processing step.

[1624] Step 2:

[1625] The server passes the query received from the user's input means to the emotion engine. The emotion engine analyzes the text data of the received query and classifies the user's emotion. Specifically, it uses IBM Watson's Tone Analyzer to determine emotions such as "anxiety," "anger," and "interest" from the input question. Emotional information such as "anxiety" is output as the analysis result.

[1626] Step 3:

[1627] The server then sends the query, along with the emotion information from the emotion engine, to the natural language processing module, which uses the Google Natural Language API to analyze the query and extract important keywords and contextual information. For example, keywords like "ChatGPT" and "security" are extracted and sent to the next processing step.

[1628] Step 4:

[1629] The server uses the extracted keywords and emotion information to search for relevant information using a database connection method. The database stores information on corporate security rules and laws. An SQL query is sent to this database to retrieve the appropriate information. For example, information on "security risks regarding the internal use of ChatGPT" is returned as a search result.

[1630] Step 5:

[1631] The server uses natural language generation to generate responses to users based on the information and emotional information obtained from the database connection means. It uses OpenAI's GPT-3 model to create prompts based on the input data and generate emotional responses. A specific example of output might be something like, "There are some risks associated with using ChatGPT, but it can be used if appropriate measures are taken."

[1632] Step 6:

[1633] The server formats the generated answer in HTML format and sends it to the user's device. The device displays the received answer on the application. The user can then decide what to do next based on the answer.

[1634] Specific actions

[1635] Users use their smartphones to enter and submit questions.

[1636] The server uses an emotion engine to analyze the user's emotions.

[1637] The server uses a natural language processing module to extract important keywords.

[1638] The server queries the database to retrieve the appropriate information.

[1639] The server uses natural language generation means to generate answers that are easy for the user to understand.

[1640] The server sends the generated response to the terminal, which displays it.

[1641] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding 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.

[1642] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with 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.

[1643] 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 robot 414.

[1644] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1645] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1646] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1647] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1648] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1649] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1650] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1651] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1652] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1653] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1654] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1655] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1656] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1657] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1658] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1659] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1660] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1661] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1662] The following is further disclosed regarding the above embodiment.

[1663] (Claim 1)

[1664] a user input means;

[1665] natural language processing means for analyzing queries from said user input means;

[1666] database connection means for searching for relevant information based on the query analyzed by the natural language processing means;

[1667] a natural language generation means for generating an answer that is easy for the user to understand based on the information acquired by the database connection means;

[1668] A system including display means for presenting to a user the answers generated by said natural language generation means.

[1669] (Claim 2)

[1670] 10. The system of claim 1, wherein the user input means receives a query from a user via a web browser.

[1671] (Claim 3)

[1672] 10. The system of claim 1, wherein the natural language processing means extracts important keywords and contextual information from a query.

[1673] (Claim 4)

[1674] 2. The system of claim 1, wherein the database connection means accesses a database of security regulations and related laws.

[1675] (Claim 5)

[1676] 2. The system of claim 1, wherein the natural language generation means generates answers in a user-friendly format.

[1677] (Claim 6)

[1678] 2. The system according to claim 1, wherein the display means displays the answer generated by the natural language generation means on the user's terminal.

[1679] (Claim 7)

[1680] 3. The system of claim 2, wherein the user input means is a web interface including a question form and a submit button.

[1681] "Example 1"

[1682] (Claim 1)

[1683] a user input means;

[1684] natural language processing means for analyzing queries from said user input means;

[1685] database connection means for searching for relevant information based on the query analyzed by the natural language processing means;

[1686] a natural language generation means for generating an answer that is easy for the user to understand based on the information acquired by the database connection means;

[1687] a display means for presenting to a user the answer generated by the natural language generation means;

[1688] the user input means receives user input via a web interface and sends a query by operating a send button;

[1689] means for analyzing keywords and contextual information in a query using a natural language processing module;

[1690] a means for generating an answer based on information acquired from a database by the natural language generation means using a natural language generation module;

[1691] The system wherein the display means includes means for receiving the generated answers and displaying them on a web interface.

[1692] (Claim 2)

[1693] 10. The system of claim 1, wherein the user input means receives a query from a user via a web browser.

[1694] (Claim 3)

[1695] 10. The system of claim 1, wherein the natural language processing means extracts keywords and contextual information from a query.

[1696] "Application Example 1"

[1697] (Claim 1)

[1698] a user input means;

[1699] natural language processing means for analyzing queries from said user input means;

[1700] database connection means for searching for relevant information based on the query analyzed by the natural language processing means;

[1701] a natural language generation means for generating an answer that is easy for the user to understand based on the information acquired by the database connection means;

[1702] a display means for presenting to a user the answer generated by the natural language generation means;

[1703] means for retrieving relevant laws and guidelines from a database using keywords extracted from the query and incorporating them into the generated answer;

[1704] A means for generating guidelines based on input information using a generative AI model;

[1705] A system including:

[1706] (Claim 2)

[1707] 10. The system of claim 1, wherein the user input means receives a query from a user via a smartphone application.

[1708] (Claim 3)

[1709] 2. The system of claim 1, wherein the generative AI model uses prompt sentences to generate guidelines for businesses to use the generative AI safely.

[1710] "Example 2: Combining Emotion Engines"

[1711] (Claim 1)

[1712] a user input means;

[1713] natural language processing means for analyzing queries from said user input means;

[1714] sentiment analysis means for analyzing the query and the sentiment of the user;

[1715] a database connection means for retrieving relevant information based on the query and sentiment analyzed by the natural language processing means and sentiment analysis means;

[1716] a natural language generation means for generating an answer that is easy for the user to understand based on the information acquired by the database connection means and the user's feelings;

[1717] A system including display means for presenting to a user the answers generated by said natural language generation means.

[1718] (Claim 2)

[1719] 10. The system of claim 1, wherein the user input means receives queries from a user via a web interface.

[1720] (Claim 3)

[1721] 10. The system of claim 1, wherein the natural language processing means extracts important keywords and contextual information from a query.

[1722] "Application example 2 when combining emotion engines"

[1723] (Claim 1)

[1724] a user input means;

[1725] natural language processing means for analyzing queries from said user input means;

[1726] database connection means for searching for relevant information based on the query analyzed by the natural language processing means;

[1727] natural language generation means for generating an answer that takes into consideration the user's feelings based on the information acquired by the database connection means;

[1728] a display means for presenting to a user the answer generated by the natural language generation means;

[1729] The system includes an emotion engine for analyzing emotions of the user.

[1730] (Claim 2)

[1731] 10. The system of claim 1, wherein the user input means receives a query from a user via a web browser.

[1732] (Claim 3)

[1733] 10. The system of claim 1, wherein the natural language processing means extracts important keywords and contextual information from a query. [Explanation of symbols]

[1734] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. a user input means; natural language processing means for analyzing queries from said user input means; database connection means for searching for relevant information based on the query analyzed by the natural language processing means; a natural language generation means for generating an answer that is easy for the user to understand based on the information acquired by the database connection means; A system including display means for presenting to a user the answers generated by said natural language generation means.

2. 2. The system of claim 1, wherein the user input means receives queries from a user via a web browser.

3. The system of claim 1 , wherein the natural language processing means extracts important keywords and contextual information from a query.

4. 2. The system of claim 1, wherein said database connection means accesses a database relating to security regulations and related laws and regulations.

5. 2. The system of claim 1, wherein the natural language generator generates answers in a user-friendly format.

6. 2. The system according to claim 1, wherein said display means displays the answer generated by said natural language generation means on a terminal of the user.

7. 3. The system of claim 2, wherein the user input means is a web interface including a question form and a submit button.

Images