System

Abstract

To provide a system.SOLUTION: A system includes: means for acquiring a prompt text transmitted from a terminal device; means for transmitting the prompt text to a plurality of different generative AI models; means for causing each of the generative AI models to acquire a response generated based on the prompt text; and means for causing each of the generative AI models to transmit the response to the terminal device.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] With conventional AI systems, it was difficult to obtain answers to a single prompt from multiple AIs, and it was also difficult to create a new prompt based on the multiple answers obtained and send it again to the AI. Furthermore, there was no way to determine the credibility of the same answer, and there was no system that could handle chat-type or image-generating AI. [Means for solving the problem]

[0005] The system of the present invention includes a means for sending the same prompt content to multiple AIs, a means for receiving answers from multiple AIs, a means for creating and sending a new prompt based on the multiple received answers, a means for determining the credibility of the same answer content, and a means for supporting chat-based and image-generating AIs. This makes it possible to obtain answers from multiple AIs, and also to create a new prompt based on the obtained answers and send it again to the AI. Furthermore, it is possible to provide a system that can determine the credibility of the same answer content and is compatible with chat-based and image-generating AIs. [Brief explanation of the drawings]

[0006] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [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. 2 is a sequence diagram showing a flow of processing in the data processing system according to the first embodiment of the first form example. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Embodiment 1. [Figure 13] FIG. 10 is a sequence diagram showing a processing flow of a data processing system in a second embodiment of the second form example. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Embodiment Example 2. [Figure 15] FIG. 10 is a sequence diagram showing the flow of processing in a data processing system according to a third embodiment of the third embodiment. [Figure 16] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Embodiment 3. [Figure 17] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the first embodiment of the first form example when an emotion engine is combined. [Figure 18] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Form Example 1 when an emotion engine is combined. [Figure 19] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the second embodiment of the second form example when an emotion engine is combined. [Figure 20] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Form Example 2 when an emotion engine is combined. [Figure 21] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the third embodiment of the third form example when an emotion engine is combined. [Figure 22] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Form Example 3 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0007] 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.

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

[0009] 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, the 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), an APU (Accelerated Processing Unit), or a TPU (TENSOR PROCESSING UNIT (registered trademark)).

[0010] 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.

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

[0012] In the following embodiments, a communication I / F (Interface) with a 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.

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. 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."

[0014] [First embodiment]

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

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

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

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

[0019] The reception device 38 includes a touch panel 38A, 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.

[0020] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (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.

[0021] 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.

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

[0023] 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.

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

[0025] In the smart device 14, the processor 46 performs the reception output 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.

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

[0027] "Example 1"

[0028] The system of the present invention has a function to send the same prompt content to multiple AIs via API. Specifically, the system sends the same prompt to AIs such as OpenAI (registered trademark) via Azure (registered trademark) and Bard (registered trademark) via GCP.

[0029] "Example 2"

[0030] In addition, as a means of receiving answers from multiple AIs, it has the function of receiving answers from each AI via API and managing them centrally. Specifically, it receives answers from each AI and stores them in a database.

[0031] "Example 3"

[0032] Furthermore, as a means of creating and sending new prompts based on multiple received answers, the system has the ability to analyze the received answers, generate new prompts based on them, and send them to the AI. Specifically, it extracts keywords from the received answers and generates new prompts based on them.

[0033] "Example 4"

[0034] In addition, as a means of determining the credibility of the same answer, the system has a function to compare answers received from multiple AIs and, if there are answers with the same content, rate the credibility of that answer highly. Specifically, the system compares answers received from multiple AIs and, if there are answers with the same content, rates the credibility of that answer highly.

[0035] "Example 5"

[0036] Furthermore, as a means of supporting both chat and image generation AI, it has an interface that supports both chat and image generation AI. Specifically, it has the function of receiving both text responses from chat AI and image responses from image generation AI and displaying them appropriately.

[0037] The processing flow of each embodiment will be described below.

[0038] "Example 1"

[0039] Step 1: User enters prompt.

[0040] Step 2: The system sends identical prompts to each AI via API (e.g., OpenAI via Azure, Bard via GCP).

[0041] Step 3: The system receives the answers from each AI via API.

[0042] Step 4: The system centralizes the responses received and stores them in a database.

[0043] "Example 2"

[0044] Step 1: The system retrieves the answers from each AI stored in the database.

[0045] Step 2: The system analyzes the answers obtained and extracts keywords.

[0046] Step 3: The system generates a new prompt based on the extracted keywords.

[0047] Step 4: Send the new system-generated prompts to each AI.

[0048] "Example 3"

[0049] Step 1: The system compares the answers it receives from multiple AIs.

[0050] Step 2: If there is an answer with the same content, the system will rate the credibility of that answer highly.

[0051] Step 1: The system receives both a text response from the chat AI and an image response from the image generation AI.

[0052] Step 2: The system displays the received text and image answers appropriately.

[0053] Example 1

[0054] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0055] In conventional systems, it was difficult to send the same prompt to multiple generative AI models and compare their responses. Furthermore, there was a lack of means to determine the credibility of the responses from each generative AI model, which sometimes raised doubts about the reliability of the information users received. Furthermore, there was no general-purpose system that could handle chat-based and image-based generative AI models.

[0056] 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.

[0057] In this invention, the server includes a means for a user to input a prompt text, a means for a terminal to send the prompt text to the server, a means for the server to receive the prompt text, a means for the server to send the prompt text to a generative AI model via a cloud service, a means for the server to receive responses from each generative AI model, a means for the server to return the response to the user, a means for sending the same prompt content to multiple generative AI models, a means for receiving answers from multiple generative AI models, a means for creating and sending a new prompt based on the multiple received answers, a means for determining the credibility of the same answer content, and a means for supporting chat-based and image-based generative AI models. This makes it possible to compare responses from multiple generative AI models and determine their credibility. Furthermore, a general-purpose system compatible with chat-based and image-based generative AI models can be provided.

[0058] "User" means an individual or entity that utilizes the system to enter prompts and receive responses from the generative AI model.

[0059] A "terminal" is a device on which a user inputs a prompt and sends it to a server, and includes a PC, smartphone, tablet, etc.

[0060] The "server" is a computer system whose role is to receive prompts, send them to the generative AI model, receive responses, and return them to the user.

[0061] A "prompt" is a question or instruction text that a user enters into a generative AI model.

[0062] "Cloud services" is a general term for computing resources and services provided over the internet, and is a platform that enables access to generative AI models.

[0063] A "generative AI model" is an artificial intelligence algorithm or system that generates a response to a prompt from a user.

[0064] A "response" is an answer or piece of information that a generative AI model generates based on a prompt.

[0065] "Credibility" is a criterion for assessing whether the responses from a generative AI model are accurate and trustworthy.

[0066] A "chat-based generative AI model" is a generative AI model designed for text-based dialogue.

[0067] An "image generation system generative AI model" is a generative AI model designed to generate images based on prompt text.

[0068] A "new prompt" is a new question or instruction text that is created based on multiple responses received.

[0069] This invention is a system that sends the same prompt sentence to multiple generative AI models and compares their responses. A specific embodiment of this system is described below.

[0070] The user inputs a prompt using a terminal, which can be a PC, smartphone, tablet, or other device, and sends the input prompt to the server. The server receives the prompt and sends it to the generative AI model via a cloud service.

[0071] The server uses a platform commonly used for cloud services, such as Azure or GCP, to access the generative AI model. Specifically, it uses the Azure SDK or GCP SDK to generate an API request and includes the prompt statement as the payload.

[0072] The server receives responses from each generative AI model and returns them to the user. Responses are answers or information generated by the generative AI model based on the prompt. The server compiles the received responses and returns them to the device, which then displays them to the user.

[0073] As a specific example, if a user enters a prompt such as "What's the weather like today?", the server sends this prompt to multiple generative AI models via a cloud service. For example, the server sends the same prompt to a generative AI model provided via Azure and a generative AI model provided via GCP. After receiving responses from each generative AI model, the server returns these responses to the user. The user will see the message "Generative AI model A: Today's weather is sunny. Generative AI model B: Today's weather is cloudy" displayed on their device screen.

[0074] This system allows users to compare responses from multiple generative AI models and judge their credibility. It also provides a general-purpose system that supports chat and image generation generative AI models.

[0075] Examples of prompts include:

[0076] "What's the weather like today?"

[0077] "Please tell me the latest news."

[0078] How do I write a Hello World program in Python?

[0079] In this way, users can ask a variety of questions and give instructions to the generative AI model and use the responses it receives.

[0080] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0081] Step 1:

[0082] The user enters a prompt statement.

[0083] Specifically, the user enters "What's the weather like today?" into the input field of the terminal. The entered prompt text is temporarily stored in the terminal's memory. Input: The prompt text entered by the user. Output: The prompt text stored in the terminal.

[0084] Step 2:

[0085] The terminal sends the prompt to the server.

[0086] The device uses the JavaScript (registered trademark) fetch API to send an HTTP POST request containing the prompt text to the server. The request also includes the user's identification information and a timestamp. Input: The prompt text saved on the device. Output: The HTTP request sent to the server.

[0087] Step 3:

[0088] The server receives the prompt.

[0089] The server uses the Node.js Express framework to extract the prompt text from the received HTTP request. The extracted prompt text is stored in the server's memory. Input: HTTP request sent to the server. Output: Prompt text stored on the server.

[0090] Step 4:

[0091] The server sends the prompt to the generative AI model via a cloud service.

[0092] The server uses the Azure SDK to send a prompt to the API of Generative AI Model A. At the same time, it uses the GCP SDK to send the same prompt to the API of Generative AI Model B. Input: Prompt text saved on the server. Output: Prompt text sent to Generative AI Model A and Generative AI Model B.

[0093] Step 5:

[0094] The server receives the response from each generative AI model.

[0095] The server waits for HTTP responses from Generative AI Model A and Generative AI Model B, and receives each response. The received responses are stored in the server's memory. Input: HTTP responses from Generative AI Model A and Generative AI Model B. Output: Responses stored on the server.

[0096] Step 6:

[0097] The server returns the response to the user.

[0098] The server compiles the received responses in JSON format and returns them to the device. The device then displays this to the user. Specifically, the browser screen displays "Generated AI model A: Today's weather is sunny. Generated AI model B: Today's weather is cloudy." Input: Response stored on the server. Output: Response displayed to the user.

[0099] (Application example 1)

[0100] Next, a description will be given of Application Example 1 of Embodiment 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."

[0101] Conventional systems using generative AI models lacked a means to efficiently compare and integrate responses from multiple AIs, making it difficult to provide information to users in an optimal format. Furthermore, the lack of consideration for implementation as an application installed on a specific device resulted in poor user convenience. This posed a challenge, limiting the scope of use of generative AI models.

[0102] 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.

[0103] In this invention, the server includes means for sending the same prompt content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending new prompts based on the received multiple answers, means for determining the credibility of the same answer content, means for supporting chat-type and image generation-type AIs, means for comparing and integrating the received multiple answers and providing them to the user in an optimal form, and means including an application to be installed on a smartphone. This makes it possible to efficiently compare and integrate answers from multiple generative AI models and provide information to the user in an optimal form.

[0104] "Artificial intelligence" refers to computer systems that have the ability to analyze data, learn, and make decisions.

[0105] A "prompt" is an input sentence that causes an artificial intelligence to perform a specific task.

[0106] An "answer" is the output generated by an artificial intelligence in response to a prompt.

[0107] "Credibility" is a criterion for evaluating the accuracy and reliability of an answer.

[0108] "Chat AI" is AI for conducting text-based conversations.

[0109] "Image generation artificial intelligence" is artificial intelligence that has the ability to generate images.

[0110] "Comparison and synthesis" is the process of evaluating multiple responses and combining them into one optimal answer.

[0111] A "smartphone" is a mobile phone that can connect to the Internet and run applications.

[0112] An "application" is a software program designed to perform a specific task.

[0113] A system for implementing this invention sends the same prompt content to multiple artificial intelligences, receives responses from the multiple artificial intelligences, compares and integrates them, and provides them to a user. The system is implemented as an application installed on a smartphone.

[0114] Hardware and software used

[0115] Hardware: Smartphone

[0116] Software: Python, cloud service APIs (Azure's OpenAI API, GCP's Bard API)

[0117] System Operation

[0118] 1. Entering a prompt: The user enters a prompt through a smartphone application, for example, "Tell me about the city of the future."

[0119] 2. Sending prompts: The server sends the input prompt to OpenAI via Azure and Bard via GCP, allowing the same prompt to be sent to multiple AIs.

[0120] 3. Receiving answers: The server receives answers from each AI. For example, OpenAI answers "The city of the future is a place where smart technology and sustainable energy solutions come together," while Bard answers "The city of the future is an environment where AI and IoT are highly integrated."

[0121] 4. Comparing and integrating answers: The server compares the multiple answers received and integrates them in the most optimal way. For example, it might integrate them as "The city of the future is an environment where smart technology and sustainable energy solutions are combined, and AI and IoT are highly integrated."

[0122] 5. Providing to users: The integrated answers are provided to users through a smartphone application, allowing them to obtain information from multiple AIs at once.

[0123] Specific examples

[0124] If the user types the prompt "Tell me about the city of the future," the system will:

[0125] Example prompt: Tell me about the city of the future.

[0126] Based on this prompt, OpenAI via Azure and Bard via GCP provide their respective answers, which are then integrated and presented to the user. For example, the integrated answer looks like this:

[0127] Example of an integrated answer: The city of the future is a blend of smart technology and sustainable energy solutions, with highly integrated AI and IoT.

[0128] In this way, users can efficiently obtain information from multiple generative AI models and utilize it in the most optimal way.

[0129] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0130] Step 1:

[0131] A user starts an application on their smartphone and inputs a prompt. The input prompt is text data such as "Tell me about the city of the future." The input prompt is then sent to the server.

[0132] Step 2:

[0133] The server sends the received prompt sentence to OpenAI via Azure and Bard via GCP. Specifically, it sends the prompt sentence using the API of each cloud service. The input is the prompt sentence, and the output is the answer from each AI.

[0134] Step 3:

[0135] The server receives answers from each AI. For example, OpenAI returns text data such as "The city of the future is a place where smart technology and sustainable energy solutions are combined," and Bard returns text data such as "The city of the future is an environment where AI and IoT are highly integrated." The input is the answer from each AI, and the output is a list of the received answers.

[0136] Step 4:

[0137] The server compares and integrates the multiple responses it receives. Specifically, it analyzes the content of each response and integrates them while eliminating duplicates. For example, it might integrate the following statement: "The city of the future is an environment where smart technology and sustainable energy solutions are combined, and AI and IoT are highly integrated." The input is a list of received responses, and the output is the integrated response.

[0138] Step 5:

[0139] The server sends the integrated answer to a smartphone application, whose input is the integrated answer and whose output is text data displayed on the user's smartphone.

[0140] Step 6:

[0141] The user can view the integrated answer through a smartphone application, which allows the user to obtain information from multiple generative AI models at once. The input is the integrated answer sent from the server, and the output is text data that the user can view.

[0142] Example 2

[0143] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0144] In conventional AI systems, it was difficult to centrally manage responses from multiple AIs and determine their credibility. Furthermore, the system lacked the ability to efficiently store received responses and search and view them later. This made it time-consuming for users to compare and consider responses from multiple AIs, hindering efficient use.

[0145] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0146] In this invention, the server includes means for sending the content of the same input sentence to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending a new input sentence based on the multiple received answers, means for determining the credibility of the same answer content, means for handling chat-type and image-generating AIs, means for centrally managing answers from each AI, means for saving received answers in a database, and means for searching and viewing saved answers. This allows answers from multiple AIs to be efficiently managed, their credibility to be determined, and users to easily search and view them.

[0147] "Artificial intelligence" refers to computer systems that have the ability to learn from data and perform specific tasks automatically.

[0148] An "input sentence" is a textual instruction or question given by a user to an artificial intelligence.

[0149] An "answer" is a text or other form of output that an artificial intelligence generates based on an input sentence.

[0150] "Credibility" is a criterion for assessing whether answers from artificial intelligence are accurate and trustworthy.

[0151] "Chat AI" is AI designed for text-based conversations.

[0152] "Image generation artificial intelligence" is artificial intelligence that has the ability to generate images based on input text and other data.

[0153] "Centralized management" means integrating and managing multiple pieces of information and data in one system or location.

[0154] A "database" is a system for efficiently storing, searching, and managing data.

[0155] "Searching and viewing" means searching for stored data based on specific conditions and checking its contents.

[0156] The present invention relates to a system for centrally managing responses from multiple artificial intelligences and determining their credibility. Specific embodiments of this system will be described below.

[0157] The server sends the prompt text entered by the user to multiple generative AI models. Generative AI models include, for example, models specialized in text generation and models specialized in image generation. Specific generative AI models include OpenAI's GPT-3 (registered trademark) and Google's BERT (registered trademark).

[0158] The server receives the prompt text entered by the user and sends an API request to each generative AI model. The API request is made using an HTTP POST request, sending data including the prompt text. The answer from each generative AI model is returned in JSON format.

[0159] The server receives the answers from each generative AI model and stores them in a database. A relational database such as MySQL (registered trademark) or PostgreSQL is used as the database. The stored data includes prompt sentences and the corresponding answers from each generative AI model.

[0160] Users can search and view past prompts and their responses through the system interface. For example, they can display a list of responses from each generative AI model to the prompt, "Tell me about the future of AI-based text generation."

[0161] For example, if a user enters the prompt "Please tell me about the future of AI text generation," the server sends this prompt to each generative AI model. The server receives the answers from each generative AI model and stores them in a database. The user can later compare the answers of each generative AI model to this prompt.

[0162] Examples of prompts include:

[0163] "Please tell me about the future of AI text generation."

[0164] "Please explain the effects of climate change."

[0165] "Tell me about the latest technology trends"

[0166] This system efficiently manages answers from multiple generative AI models, determines their credibility, and allows users to easily search and view them.

[0167] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0168] Step 1:

[0169] The user enters a prompt statement.

[0170] The user enters a prompt through the system interface, for example, by typing "Tell me about the future of AI-generated text" into a text box on a web browser and clicking the submit button.

[0171] Input: The prompt text entered by the user.

[0172] Output: The prompt text sent to the server.

[0173] Step 2:

[0174] The server sends the prompt sentence to each generative AI model.

[0175] The server receives the prompt sentence entered by the user and sends an API request to each generative AI model. Specifically, it sends the prompt sentence to the API endpoint of the generative AI model using an HTTP POST request.

[0176] Input: The prompt received from the user.

[0177] Output: API requests sent to each generative AI model.

[0178] Step 3:

[0179] The server receives the answers from each generative AI model.

[0180] The server receives API responses from each generative AI model. Responses are usually returned in JSON format. For example, from GPT-3, the server receives the answer, "The future of AI text generation is very bright."

[0181] Input: API response from each generative AI model.

[0182] Output: The response data received by the server.

[0183] Step 4:

[0184] The server stores the received response in a database.

[0185] The server saves the received answers in a database. Specifically, it uses a relational database such as MySQL or PostgreSQL to store the prompt sentences and the corresponding answers of each generative AI model in a table.

[0186] Input: The answer data received from each generative AI model.

[0187] Output: Response data stored in a database.

[0188] Step 5:

[0189] Users can search and view saved answers.

[0190] Users can search and view past prompts and their responses through the system interface. For example, the system displays a list of responses from each generative AI model to the prompt, "Tell me about the future of AI-based text generation."

[0191] Input: The user's search query.

[0192] Output: The prompt and answer data displayed as search results.

[0193] (Application example 2)

[0194] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0195] Conventional AI systems have difficulty managing responses from multiple AIs in a unified manner and providing users with the most appropriate answer. Furthermore, they lacked the means to determine the credibility of the answers, making it impossible to provide users with highly reliable information. Furthermore, they lacked the ability to save answers in a database and refer to them later, making efficient information management impossible.

[0196] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0197] In this invention, the server includes means for sending the same input content to multiple AIs, means for receiving answers from the multiple AIs, means for creating and sending new input content based on the received multiple answers, means for determining the credibility of the same answer content, means for handling chat-type and image-generating AIs, means for saving answers from the multiple AIs in a database, means for selecting the optimal answer from the saved answers, and means for providing the selected optimal answer to the user. This makes it possible to centrally manage answers from the multiple AIs and provide the user with a highly reliable optimal answer.

[0198] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and learn, reason, and self-correct.

[0199] "Input" refers to questions or instructions provided by users to the artificial intelligence.

[0200] An "answer" is a response or information generated by artificial intelligence based on input.

[0201] A "database" is a system for efficiently storing, managing, and retrieving information.

[0202] "Credibility" is an indicator of whether information or answers are accurate and trustworthy.

[0203] "Chat AI" is AI that can conduct text-based conversations.

[0204] "Image generation artificial intelligence" is artificial intelligence that can generate images based on input content.

[0205] The "best answer" is the most appropriate and reliable answer selected from multiple answers.

[0206] "User" refers to a person or organization that uses the system to ask questions or give instructions.

[0207] "Cloud services" refer to computer resources and services provided over the Internet.

[0208] A "network" is a system in which multiple computers and devices are connected to each other and used to exchange information.

[0209] The system for carrying out this invention has the function of sending the same input content to multiple AIs, receiving responses from the multiple AIs, and managing them in a unified manner. A specific embodiment of this system will be described below.

[0210] System Configuration

[0211] The server includes the following means:

[0212] 1. Input content transmission means: Sends questions and instructions (prompt text) entered by the user to multiple artificial intelligences.

[0213] 2. Answer receiving means: Receive answers from multiple artificial intelligences.

[0214] 3. Method for creating new input content: Create new input content based on the multiple answers received and send it to the artificial intelligence again.

[0215] 4. Credibility judgment method: judge the credibility of the same answers.

[0216] 5. Database storage means: Responses from multiple artificial intelligences are stored in a database.

[0217] 6. Best answer selection method: Select the best answer from the saved answers.

[0218] 7. User-provided means: The selected best answer is provided to the user.

[0219] Hardware and software used

[0220] Hardware: Servers and user devices (smartphones, PCs, etc.) that can connect to the Internet.

[0221] Software: Python, SQLite, APIs for each artificial intelligence model (e.g., GPT-4 (registered trademark), BERT, T5, etc.).

[0222] Data processing and calculation

[0223] The server receives input from the user and sends it to multiple AIs. After receiving answers from each AI, the server stores these answers in a database. The server selects the best answer from the stored answers and provides it to the user. The credibility judgment means evaluates the credibility of the same answer content and provides highly reliable information.

[0224] Specific examples

[0225] When a user asks, "What's the weather like today?", the server sends this prompt to multiple AIs. For example, GPT-4 may give an answer like, "It's sunny today," BERT may give an answer like, "It's cloudy today," and T5 may give an answer like, "It's going to rain today." These answers are stored in a database and evaluated using a credibility assessment tool. The best answer is selected and served to the user.

[0226] Example prompt sentence:

[0227] What's the weather like today?

[0228] In this way, it becomes possible to centrally manage answers from multiple AIs and provide users with the most reliable and optimal answers.

[0229] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0230] Step 1:

[0231] The user uses the terminal to enter input (prompt text).

[0232] Input: The prompt text entered by the user (e.g., "What's the weather like today?").

[0233] Output: The input is sent to the server.

[0234] Specific operation: When the user enters a question in the input field of the terminal and presses the send button, the prompt sentence is sent to the server.

[0235] Step 2:

[0236] The server receives the input and sends it to multiple generative AI models.

[0237] Input: The prompt text sent by the user.

[0238] Output: The prompt sentence is sent to multiple generative AI models.

[0239] Specific operation: The server sends the received prompt sentence to the API endpoint of each generative AI model (GPT-4, BERT, T5, etc.).

[0240] Step 3:

[0241] The server receives the answers from each generative AI model.

[0242] Input: The prompt sent to each generative AI model.

[0243] Output: The answer from each generative AI model.

[0244] Specific operation: Each generative AI model generates an answer based on the prompt sentence and sends the answer back to the server.

[0245] Step 4:

[0246] The server stores the received response in a database.

[0247] Input: The answers from each generative AI model.

[0248] Output: The answer is saved in the database.

[0249] Specific operation: The server stores the received answer in an SQLite database, along with information to identify which generative AI model the answer came from.

[0250] Step 5:

[0251] The server selects the best answer from the stored answers.

[0252] Input: Multiple answers stored in a database.

[0253] Output: The best answer.

[0254] What it does: The server analyzes the stored answers and runs an algorithm to select the best answer based on the authenticity and content of the answer.

[0255] Step 6:

[0256] The server provides the best answer to the user.

[0257] Input: Best answer.

[0258] Output: The answer provided to the user.

[0259] Specific operation: The server sends the selected optimal answer to the user's device and displays it on the user's screen.

[0260] In this way, it is possible to centrally manage responses from multiple generative AI models to prompts entered by users, and provide highly reliable and optimal answers.

[0261] Example 3

[0262] Next, a description will be given of a third embodiment of the third embodiment. 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."

[0263] Conventional systems have difficulty effectively analyzing responses from multiple generative AI models and generating new prompts. Furthermore, they lack a means to determine the credibility of responses from multiple generative AI models, making it impossible to provide users with reliable information. Furthermore, there was no unified interface for chat and image generation generative AI models, resulting in an inconsistent user experience.

[0264] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0265] In this invention, the server includes means for receiving input from a user, means for sending prompts to multiple generative AI models, means for receiving responses from the multiple generative AI models, means for analyzing the received responses and extracting keywords, means for generating new prompts based on the extracted keywords, means for comparing responses from the multiple generative AI models and determining their credibility, and means for supporting chat-based and image-based generative AI models. This makes it possible to effectively analyze responses from multiple generative AI models based on user input and generate new prompts. Furthermore, by determining the credibility of responses from multiple generative AI models, highly reliable information can be provided to the user. Furthermore, by providing a unified interface for chat-based and image-based generative AI models, a consistent user experience can be achieved.

[0266] A "user" is an entity that inputs prompts into the system and receives answers from the generative AI model.

[0267] A "server" is a device or system that receives input from users, sends prompts to multiple generative AI models, and receives and analyzes responses.

[0268] A "generative AI model" is an artificial intelligence model that generates answers such as text and images based on a given prompt.

[0269] A "prompt" is text that contains instructions or questions that are input to a generative AI model.

[0270] An "answer" is an output result, such as text or an image, that a generative AI model generates based on a prompt.

[0271] "Keywords" are important words or phrases extracted by analyzing the answers from the generative AI model.

[0272] "Credibility" is the reliability of an answer, assessed based on the degree to which answers from multiple generative AI models match.

[0273] A "chat-based generative AI model" is a generative AI model for conducting text-based conversations.

[0274] An "image generation system generative AI model" is a generative AI model for generating images based on prompt sentences.

[0275] An "interface" is a means or method by which a user interacts with a system, and corresponds to the generative AI models of chat and image generation systems.

[0276] MODE FOR CARRYING OUT THE INVENTION

[0277] This invention relates to a system that receives input from a user, analyzes responses from multiple generative AI models, and generates and sends new prompt sentences. A specific embodiment of this system is described below.

[0278] Hardware and software used

[0279] The server uses the following hardware and software:

[0280] Hardware: A server machine with a powerful processor, sufficient memory, and a fast network connection.

[0281] Software: Operating systems (e.g., Linux), programming languages ​​(e.g., Python), natural language processing libraries (e.g., spaCy, NLTK), text matching algorithms (e.g., difflib, fuzzywuzzy), generative AI models (e.g., OpenAI's GPT-3, DALL-E).

[0282] System processing flow

[0283] A user uses a terminal to input a prompt to the system, such as "Generate an image of a cat" or "What is the diameter of the Earth?"

[0284] The server receives input from the user and sends prompts to multiple generative AI models using the APIs of the generative AI models (e.g., OpenAI's ChatGPT (registered trademark) or DALL-E).

[0285] The server receives answers from each generative AI model, which may be in text or image format.

[0286] The server analyzes the received responses and extracts keywords using Python's natural language processing libraries, such as spaCy and NLTK. For example, in response to a prompt such as "Generate an image of a cat," the server extracts keywords such as "cat" and "image."

[0287] The server generates a new prompt based on the extracted keywords. This prompt is generated using a generative AI model (e.g., OpenAI's GPT-3). For example, it generates a new prompt such as "Generate a more detailed image of a cat."

[0288] The server compares answers from multiple generative AI models and, if there are identical answers, rates them as more reliable. This comparison uses text matching algorithms like "difflib" and "fuzzywuzzy."

[0289] The server sends the most credible answer to the terminal for display to the user, and the terminal shows the displayed text answer and image answer to the user.

[0290] Specific examples

[0291] Example 1: Receiving input from the user

[0292] A user types "generate an image of a cat" into a terminal.

[0293] The server receives this input and logs it.

[0294] Example 2: Sending prompts to multiple AIs

[0295] The server sends a prompt to ChatGPT and DALL-E APIs: "Generate a cat image."

[0296] The server checks the status of the API request and proceeds to the next step if successful.

[0297] Example 3: Receiving answers from AI

[0298] The server receives text responses from ChatGPT and image responses from DALL-E.

[0299] The server temporarily stores the received response.

[0300] Example 4: Answer analysis and keyword extraction

[0301] The server analyzes ChatGPT's text responses using "spaCy" and extracts keywords such as "cat" and "image."

[0302] The server records the extracted keywords in a log.

[0303] Example 5: Creating a new prompt

[0304] Based on the extracted keywords, the server generates a new prompt such as "Generate a more detailed image of a cat."

[0305] The server sends this prompt to the generative AI model.

[0306] Example 6: Judging credibility

[0307] The server uses "fuzzywuzzy" to compare answers received from multiple generative AI models, and if there are answers with the same content, it rates their credibility highly.

[0308] The server logs the results of the credibility assessment.

[0309] Example 7: Displaying results to the user

[0310] The server sends the most credible answer to the terminal for display to the user.

[0311] The terminal shows the displayed text and image answers to the user.

[0312] In this way, this system can effectively analyze responses from multiple generative AI models based on user input and generate new prompts. Furthermore, by determining the credibility of responses from multiple generative AI models, it is possible to provide highly reliable information to the user. Furthermore, by providing a unified interface for chat-based and image-based generative AI models, it is possible to ensure a consistent user experience. The flow of the identification process in Example 3 is described with reference to FIG. 15.

[0313] Step 1:

[0314] A user uses a terminal to input a prompt to the system, such as "Generate an image of a cat" or "What is the diameter of the Earth?"

[0315] Input: The prompt text entered by the user.

[0316] Output: The prompt received by the server.

[0317] Specific behavior: The user enters a prompt into the input field of the terminal and presses the send button. The server receives this input and records it in the log.

[0318] Step 2:

[0319] The server sends the received prompts to multiple generative AI models using the APIs of the generative AI models (e.g., OpenAI's ChatGPT or DALL-E).

[0320] Input: The prompt received by the server.

[0321] Output: The prompt sent to each generative AI model.

[0322] Specific operation: The server sends the prompt sentence as an API request to the generative AI model, checks the status of the API request, and proceeds to the next step if it is successful.

[0323] Step 3:

[0324] The server receives answers from each generative AI model, which may be in text or image format.

[0325] Input: The answer from each generative AI model.

[0326] Output: The answer stored on the server.

[0327] Specific operation: The server receives the API response from the generative AI model and temporarily stores the text and image answers.

[0328] Step 4:

[0329] The server analyzes the received responses and extracts keywords using Python's natural language processing libraries, spaCy and NLTK.

[0330] Input: The received text response.

[0331] Output: Extracted keywords.

[0332] Specific operation: The server analyzes the text response using "spaCy" and "NLTK" to extract keywords such as "cat" and "image." The extracted keywords are recorded in a log.

[0333] Step 5:

[0334] The server generates a new prompt based on the extracted keywords, using a generative AI model (e.g., OpenAI's GPT-3).

[0335] Input: Extracted keywords.

[0336] Output: The new prompt text.

[0337] Specific operation: The server generates a new prompt sentence based on the extracted keywords, such as "Generate a more detailed image of a cat," and sends this prompt sentence to the generative AI model.

[0338] Step 6:

[0339] The server compares answers from multiple generative AI models and, if there are identical answers, rates them as more reliable. This comparison uses text matching algorithms like "difflib" and "fuzzywuzzy."

[0340] Input: Answers from multiple generative AI models.

[0341] Output: The credibility assessment result.

[0342] Specific behavior: The server compares the received answers using "difflib" or "fuzzywuzzy", and if there are answers with the same content, rates the answer as more reliable. The server records the result of the reliability evaluation in a log.

[0343] Step 7:

[0344] The server sends the most credible answers to the terminal for display to the user, and the terminal shows the displayed text answers and image answers to the user.

[0345] Input: A credible answer.

[0346] Output: The answer that is displayed to the user.

[0347] Specific operation: The server sends a credible answer to the device, which displays it to the user. The user checks the displayed text and image answers.

[0348] (Application example 3)

[0349] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[0350] Conventional advertising material generation systems make it difficult for advertising creators to efficiently generate high-quality advertising copy and images. Furthermore, there was a lack of means to evaluate the credibility of responses from multiple AIs, which led to problems with the quality and reliability of the generated advertising materials. Furthermore, there was no interface compatible with both conversational AI and image-generating AI, making the advertising material generation process cumbersome.

[0351] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means. In this invention, the server includes means for sending the same instruction content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending new instructions based on the multiple received answers, means for determining the credibility of the same answer content, means for supporting dialogue-based AI and image-generating AI, means for generating instructions for generating advertising material, means for acquiring advertising copy and images from multiple generative AI models, means for evaluating the credibility of the acquired advertising copy and images, and means for displaying the acquired advertising copy and images. This enables advertising creators to efficiently generate high-quality, reliable advertising material.

[0352] "Multiple artificial intelligences" refers to multiple AI systems that operate using different algorithms and platforms.

[0353] "Instructions" refer to input information or commands that instruct the artificial intelligence to perform a specific task.

[0354] An "answer" is the output result generated by the artificial intelligence based on the instructions.

[0355] "Credibility" is the standard by which we evaluate whether the answers we receive are accurate and trustworthy.

[0356] "Conversational artificial intelligence" refers to AI systems that have the ability to converse with humans through text or voice.

[0357] "Image-generating artificial intelligence" refers to an AI system that has the ability to generate images based on input instructions.

[0358] "Advertising Materials" means content such as text and images used in an advertising campaign.

[0359] A "generative artificial intelligence model" is an AI model that generates content such as text and images for a specific task.

[0360] "Measures for assessing credibility" refers to methods or algorithms for assessing the accuracy and reliability of answers received from multiple AIs.

[0361] "Displaying means" refers to a method or device for visually presenting the generated advertising copy and images to the user.

[0362] The system for implementing this invention is a tool that allows advertising creators to efficiently generate high-quality advertising materials. The system has the function of sending the same instructions to multiple AIs, receiving responses from the multiple AIs, and creating and sending new instructions based on those responses. It also judges the credibility of the same responses and has an interface that supports dialogue-based AI and image-generating AI.

[0363] Hardware and software used

[0364] Hardware:

[0365] Smartphone

[0366] software:

[0367] Python

[0368] OpenAI API

[0369] PIL (Python Imaging Library)

[0370] requests

[0371] Data processing and calculation

[0372] Server Action:

[0373] 1. Prompt generation: Based on the keywords and themes entered by the ad creator, a prompt is generated to be sent to the generative AI model. For example, if the keyword "new smartphone release" is entered, the prompt will be as follows:

[0374] Ad theme: New smartphone launch. Generate the best copy and images.

[0375] 2. Obtaining answers from AI: Using OpenAI's API, we obtain ad copy and images from the generative AI model. Text answers are obtained from the conversational AI, and image answers are obtained from the image-generating AI.

[0376] 3. Credibility assessment: Compare the responses received to see if any of them are identical, thereby assessing their credibility.

[0377] 4. Displaying the results: Displaying the generated ad copy and fetching and displaying images. For example, if the generated ad copy is "This new smartphone is equipped with the latest technology" and the generated image is a high-resolution image of a smartphone, they will be visually presented to the ad creator.

[0378] Specific examples

[0379] When an ad creator enters the keyword "new smartphone release," the system generates the following prompt and sends it to the AI:

[0380] Ad theme: New smartphone launch. Generate the best copy and images.

[0381] The AI ​​generates the ad copy, "This new smartphone is equipped with the latest technology," along with a high-resolution image of the smartphone. The system evaluates these responses and, if deemed credible, displays them to the ad creator.

[0382] In this way, ad creators can efficiently generate high-quality, reliable ad materials.

[0383] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0384] Step 1:

[0385] The user inputs keywords or themes for generating advertising materials. For example, the input keyword might be "new smartphone release." This input becomes the basis for generating prompts.

[0386] Step 2:

[0387] The server generates a prompt based on the input keywords. Specifically, it creates the prompt "Advertisement theme: New smartphone release. Please generate the best ad copy and image." This prompt is used in the next step.

[0388] Step 3:

[0389] The server sends the generated prompt sentence to multiple generative AI models. In this case, the prompt sentence is sent to both conversational AI and image generation AI. The input is the prompt sentence, and the output is the answer from each AI model.

[0390] Step 4:

[0391] The server receives responses from multiple generative AI models. The conversational AI model returns ad copy, while the image-generating AI model returns ad images. For example, a high-resolution image of a smartphone might be generated along with the ad copy, "This new smartphone is equipped with the latest technology."

[0392] Step 5:

[0393] The server evaluates the credibility of the received answers. Specifically, it compares answers from multiple AIs to see if there are any answers with the same content. Answers that are judged to be highly credible are used in the next step.

[0394] Step 6:

[0395] The server displays the advertisement copy and images that are judged to be highly credible to the user. The user can review the generated advertisement copy and images and make corrections or adopt them as necessary. The output is the advertisement material presented to the user.

[0396] In this way, users can efficiently generate high-quality and reliable advertising materials.

[0397] 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.

[0398] "Example 1"

[0399] One embodiment of the present invention is a system incorporating an emotion engine. This system analyzes emotions from text and voice input from a user. Specifically, if a user inputs the text "Today was fun" to the system, the emotion engine recognizes the emotion "joy" from this text. Also, if a user inputs the text "Today was hard," the emotion engine recognizes the emotion "sadness."

[0400] "Example 2"

[0401] Furthermore, the emotion engine adjusts the AI's response based on the analyzed emotions. For example, if a user enters the text "I had fun today," and the emotion "joy" is recognized, the AI ​​will respond positively, such as "That's great! What was fun about it?" On the other hand, if a user enters the text "I had a hard day today," and the emotion "sadness" is recognized, the AI ​​will respond with an empathetic response, such as "That must have been tough. Maybe talking about it will help you feel a little better."

[0402] "Example 3"

[0403] In this way, a system incorporating an emotion engine can properly recognize the user's emotions and generate AI responses accordingly, enabling more human-like communication, which makes the user's interaction with the system feel more natural and satisfying.

[0404] The processing flow of each embodiment will be described below.

[0405] "Example 1"

[0406] Step 1: The user inputs text into the system. For example, the user inputs the text "I had a great time today."

[0407] Step 2: The emotion engine analyzes the emotion from the user's text input. In this example, the emotion "joy" is recognized.

[0408] "Example 2"

[0409] Step 1: The user inputs text into the system, for example, "Today was a tough day."

[0410] Step 2: The emotion engine analyzes the emotion from the user's text input. In this example, the emotion "sadness" is recognized.

[0411] Step 3: The emotion engine adjusts the AI's response based on the analyzed emotions. In this example, it generates an empathetic response such as, "That must be tough. Maybe talking about it will help you feel a little better."

[0412] "Example 3"

[0413] Step 1: The user provides text input to the system.

[0414] Step 2: The emotion engine analyzes the emotion from the user's text input.

[0415] Step 3: The emotion engine adjusts the AI's response based on the analyzed emotions.

[0416] Step 4: The AI-generated response is returned to the user, making the user's interaction with the system feel more natural and satisfying.

[0417] Example 1

[0418] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0419] In conventional systems, it was difficult to send the same prompt to multiple AIs and collect their responses. Furthermore, the system lacked the ability to analyze the user's emotions and generate responses based on them, making it impossible to provide appropriate responses that met the user's needs. Furthermore, there was also an insufficient means to determine the credibility of responses from multiple AIs.

[0420] 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.

[0421] In this invention, the server includes means for sending the same prompt content to multiple AIs, means for receiving answers from the multiple AIs, means for creating and sending new prompts based on the received multiple answers, means for determining the credibility of the same answer content, means for analyzing emotions from user input, means for generating responses based on the analyzed emotions, and means for supporting chat-type and image-generating AIs. This makes it possible to efficiently collect responses from the multiple AIs and provide appropriate responses based on the user's emotions.

[0422] "Artificial intelligence" is the technology that enables computer systems to learn, reason, and self-correct by imitating human intelligence.

[0423] A "prompt" is an input sentence that causes an artificial intelligence to generate a particular response.

[0424] "Emotion analysis" is a technology that identifies emotions from user input and evaluates the type and intensity of those emotions.

[0425] "Credibility" is the standard by which information or responses are judged to be accurate and trustworthy.

[0426] "Chat AI" is AI designed for text-based conversations.

[0427] "Image generation AI" is AI designed to generate images based on input data.

[0428] A "platform" is the underlying hardware and software environment for running software and services.

[0429] A "network" is a system that allows multiple computers and devices to communicate with each other.

[0430] "Merchandise" refers to goods and services that are the subject of commercial activity.

[0431] This invention is a system that sends the same prompt content to multiple artificial intelligences, collects their responses, analyzes the user's emotions, and generates an appropriate response. Specific embodiments of this system will be described below.

[0432] The server is connected to a terminal for receiving prompt sentences from the user. When the user inputs a prompt sentence using the terminal, the prompt sentence is sent to the server. The server passes the received prompt sentence to an emotion engine, which analyzes the user's emotion. The emotion engine identifies emotions from the user's input text and evaluates the type and intensity of the emotion.

[0433] The server then sends prompts to multiple AIs based on the analyzed emotional information. Specifically, the server uses the API of the cloud platform to send the same prompt to different AIs. For example, the server uses the API of cloud service A to send a prompt to AI model A, and simultaneously sends the same prompt to AI model B using the API of cloud service B.

[0434] Each AI model generates a response to the received prompt. The generated response is sent to the server. The server receives responses from multiple AI models and judges the credibility of those responses. The credibility judgment takes into account the degree of similarity of the response content and past response history.

[0435] Finally, the server returns a highly credible response to the user. For example, if a user inputs a prompt sentence such as "What's the weather like today?", the server receives a response such as "It's sunny today" from AI model A of cloud service A, and a response such as "It's cloudy today" from AI model B of cloud service B. The server compares these responses and returns a highly credible response to the user.

[0436] As a concrete example, the processing when the user inputs the prompt sentence "What's the weather like today?" is shown below.

[0437] 1. A user uses a terminal to type, "What's the weather like today?"

[0438] 2. The server receives the prompt and performs emotion analysis using the emotion engine.

[0439] 3. The server uses the API of cloud service A to send a prompt to artificial intelligence model A, and simultaneously sends the same prompt to artificial intelligence model B using the API of cloud service B.

[0440] 4. AI model A responds, "It's sunny today," and AI model B responds, "It's cloudy today."

[0441] 5. The server receives these responses and determines their authenticity.

[0442] 6. The server returns a reliable response to the user.

[0443] In this way, the server can efficiently collect responses from multiple artificial intelligences and provide an appropriate response based on the user's emotions.

[0444] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0445] Step 1:

[0446] The user enters a prompt statement.

[0447] The user uses a terminal to input a prompt to the system, for example, "What's the weather like today?" The input prompt is sent from the terminal to the server.

[0448] Input: User prompt text

[0449] Output: The prompt sent to the server

[0450] Step 2:

[0451] The server receives the prompt.

[0452] The server receives the prompt sent by the user, which becomes the input data for the next processing step.

[0453] Input: Prompt sent from the terminal

[0454] Output: Received prompt text

[0455] Step 3:

[0456] The server performs emotion analysis using an emotion engine.

[0457] The server passes the received prompt sentence to the emotion engine for emotion analysis. For example, if the user enters "I had fun today," the emotion engine recognizes the emotion of "joy."

[0458] Input: Received prompt text

[0459] Output: Parsed emotion information

[0460] Step 4:

[0461] The server sends prompt sentences to multiple artificial intelligences.

[0462] The server uses the API of the cloud platform to send the same prompt to different AIs, for example, to send a prompt to AI model A using the API of cloud service A, and at the same time to send the same prompt to AI model B using the API of cloud service B.

[0463] Input: Received prompt sentence, parsed emotion information

[0464] Output: Prompt sent to each AI

[0465] Step 5:

[0466] Each AI generates a response to the prompt.

[0467] AI model A and AI model B each generate a response to the received prompt. For example, AI model A responds, "It's sunny today," and AI model B responds, "It's cloudy today."

[0468] Input: Prompt sent to each AI

[0469] Output: The generated response

[0470] Step 6:

[0471] The server receives responses from each artificial intelligence.

[0472] The server receives responses from AI model A and AI model B. This allows information from multiple AIs to be aggregated on the server.

[0473] Input: Response from each AI

[0474] Output: Response received

[0475] Step 7:

[0476] The server determines the authenticity of the response.

[0477] The server determines the authenticity of the received response, taking into account the degree of match of the response content and past response history.

[0478] Input: Received response

[0479] Output: A credible response

[0480] Step 8:

[0481] The server returns the response to the user.

[0482] The server responds to the user with a highly credible response, for example, "OpenAI's response: It's sunny today. Bard's response: It's cloudy today."

[0483] Input: A credible response

[0484] Output: The response returned to the user

[0485] (Application example 1)

[0486] Next, a description will be given of Application Example 1 of Embodiment 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."

[0487] Conventional AI systems have had difficulty analyzing user emotions and generating optimal prompts based on those emotions. It has also been difficult to send the same prompt to multiple AIs and provide optimal customer service methods and product suggestions in real time based on their responses. This has made it difficult to improve customer satisfaction in brick-and-mortar stores.

[0488] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means. In this invention, the server includes means for sending the same prompt content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending a new prompt based on the multiple received answers, means for determining the credibility of the same answer content, means for supporting chat-type and image-generation-type AIs, means for analyzing a user's emotions using an emotion analysis engine, means for generating an optimal prompt based on the analyzed emotions, and means for sending the generated prompt to multiple AIs and providing the obtained responses. This makes it possible to provide optimal customer service methods and product suggestions in real time according to the user's emotions.

[0489] "Multiple artificial intelligences" refers to multiple AI systems that operate using different algorithms and platforms.

[0490] A "prompt" is text containing instructions or questions that instruct an artificial intelligence to perform a specific task.

[0491] An "emotion analysis engine" is software that analyzes emotions from a user's text or voice input and identifies those emotions.

[0492] A "means for determining credibility" is an algorithm or mechanism for assessing the accuracy and reliability of answers obtained from multiple artificial intelligences.

[0493] "Chat AI" refers to an AI system designed for text-based dialogue.

[0494] "Image generation artificial intelligence" is an AI system that generates images based on input data.

[0495] A "means for generating optimal prompts" is an algorithm or mechanism for generating optimal instructions or questions for the artificial intelligence based on analyzed emotions and other data.

[0496] A "means for providing a response" is a mechanism for presenting answers obtained from multiple artificial intelligences to a user or system.

[0497] The system for implementing this invention sends the same prompt content to multiple artificial intelligences, and based on the responses obtained, provides optimal customer service methods and product suggestions in accordance with the user's emotions. A specific embodiment of this system will be described below.

[0498] System Configuration

[0499] The system consists of the following main components:

[0500] 1. Sentiment Analysis Engine: Software that analyzes emotions from a user's text or voice input.

[0501] 2. Prompt generation engine: Software that generates optimal prompts based on the analyzed emotions.

[0502] 3. Multiple Artificial Intelligence: Different AI systems connected via cloud services (e.g., generative AI models via cloud services).

[0503] 4. Credibility Judgment Engine: Software that evaluates the credibility of answers obtained from multiple artificial intelligences.

[0504] 5. Response Delivery Engine: Software that delivers the resulting responses to users or systems.

[0505] Hardware and Software

[0506] The system uses the following hardware and software:

[0507] Smart glasses: A device worn by store clerks to analyze conversations and behavior with customers in real time.

[0508] Cloud Services: Platforms (e.g., cloud services) for hosting sentiment analysis engines and generative AI models.

[0509] Python: A programming language for implementing the system's programs.

[0510] Data processing and calculation

[0511] 1. Sentiment analysis: The user's text or voice input is sent to the sentiment analysis engine to analyze their emotions. For example, if the user says, "Today was fun," the sentiment analysis engine will recognize the emotion of "joy."

[0512] 2. Prompt generation: Based on the analyzed sentiment, the prompt generation engine generates the most appropriate prompt. For example, it generates a prompt such as, "What is the best way to serve a user when they are happy?"

[0513] 3. Prompt sending: Send the generated prompt to multiple AIs and get responses from each AI.

[0514] 4. Credibility judgment: The credibility of the obtained response is evaluated by the credibility judgment engine.

[0515] 5. Response provision: The response provision engine provides the evaluated responses to the store staff, who then suggest optimal customer service methods and products in real time.

[0516] Specific examples

[0517] For example, if a customer says, "I had a great time today," the emotion analysis engine recognizes the emotion of "joy." The prompt generation engine generates a prompt, "Please tell me the best way to serve a customer when they are happy," and sends it to multiple AIs. The credibility judgment engine evaluates the responses obtained and provides the optimal response to the store clerk.

[0518] Prompt Sentence Examples

[0519] "What is the best way to serve a user when they are happy?"

[0520] In this way, the system can provide optimal customer service methods and product suggestions in real time based on the user's emotions.

[0521] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0522] Step 1:

[0523] The user inputs text or voice. For example, the user says, "I had fun today." The input data is sent to the server through the smart glasses.

[0524] Step 2:

[0525] The server sends the received user input data to the emotion analysis engine. The emotion analysis engine analyzes the input data and identifies the user's emotion. For example, it recognizes the emotion "joy" from the text "I had fun today." The input is the user's text data, and the output is analyzed emotion data.

[0526] Step 3:

[0527] The server uses a prompt generation engine based on the analyzed emotion data to generate an optimal prompt. For example, for the emotion of "happiness," it generates the prompt "Please tell me the best way to serve a user when they are happy." The input is emotion data, and the output is the generated prompt.

[0528] Step 4:

[0529] The server sends the generated prompts to multiple artificial intelligences, for example, multiple generative AI models connected via a cloud service. The input is the generated prompt, and the output is the response data from each artificial intelligence.

[0530] Step 5:

[0531] The server sends the response data obtained from multiple AIs to the credibility judgment engine. The credibility judgment engine evaluates the accuracy and reliability of each response. The input is the response data, and the output is the evaluated credibility data.

[0532] Step 6:

[0533] The server selects the optimal response based on the credibility data obtained from the credibility judgment engine. The selected response is sent to the smart glasses through the response providing engine and provided to the store clerk. The input is the credibility data, and the output is the optimal response data.

[0534] Step 7:

[0535] The terminal (smart glasses) displays the optimal response data received from the server to the store clerk. Based on the displayed response, the store clerk will provide the user with the optimal customer service method and product suggestions. The input is the optimal response data, and the output is the store clerk's behavior.

[0536] In this way, the system can provide optimal customer service methods and product suggestions in real time according to the user's emotions.

[0537] Example 2

[0538] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0539] Conventional AI systems have difficulty managing responses from multiple AIs in a unified manner and providing appropriate responses based on the user's emotions. Furthermore, they lacked the means to integrate AIs provided via different platforms and select highly credible answers. Furthermore, they lacked the ability to analyze the user's emotions and adjust responses based on them, limiting the user experience.

[0540] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0541] In this invention, the server includes means for sending the same instruction content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending new instructions based on the multiple received answers, means for determining the credibility of the same answer content, means for supporting conversational and image generation AIs, means for receiving user input and analyzing emotions with an emotion analysis engine, means for adjusting the AI's response based on the analyzed emotions, means for saving the received answers in a database, means for selecting an optimal response based on emotions, and means for returning a response to the user. This makes it possible to centrally manage answers from multiple AIs and provide appropriate responses according to the user's emotions.

[0542] "Artificial intelligence" is the technology that enables computer systems to learn, reason, and self-correct by imitating human intelligence.

[0543] "Instructions" are the information or instructions given to an artificial intelligence to perform a specific task or question.

[0544] An "answer" is a response or result generated by artificial intelligence based on instructions.

[0545] "Credibility" is a criterion for assessing whether a response is accurate and trustworthy.

[0546] "Conversational AI" is AI that uses natural language processing to engage in dialogue with users.

[0547] "Image generation artificial intelligence" is artificial intelligence that generates images based on input data.

[0548] An "emotion analysis engine" is software that analyzes emotions from text entered by a user and recognizes those emotions.

[0549] A "database" is a system for efficiently storing, managing, and retrieving data.

[0550] An "optimal response" is an answer that is provided in a manner that best suits the user's feelings and situation.

[0551] A "platform" is the underlying environment or infrastructure for providing software and services.

[0552] This invention is a system that centrally manages responses from multiple artificial intelligences and provides appropriate responses according to the user's emotions. Specific embodiments of this system will be described below.

[0553] The server has a means to send the same instructions to multiple AIs. This is achieved by sending API requests using the Python requests library. For example, if a user enters "I had fun today," the server sends this text as a prompt to multiple AIs.

[0554] Next, the server has a means to receive answers from multiple AIs. This is also realized by receiving API responses using Python's requests library. For example, it receives a response from AI1 saying, "That's great! What was fun about it?"

[0555] The server also includes a means for generating and sending new instructions based on the received responses. This means includes logic for analyzing the received responses and generating the next prompt. For example, if the user inputs "Today was a bad day," the server generates a new prompt that reads, "The user is feeling sad. Please respond empathetically."

[0556] The server also includes a means for determining the authenticity of the content of the identical answers, the means including an algorithm for comparing the content of the received answers and assessing the authenticity of matching answers.

[0557] Furthermore, the server has the means to support conversational and image generation artificial intelligence, which enables not only text-based conversation but also image generation.

[0558] The server also includes a means for receiving user input and analyzing emotions using a sentiment analysis engine. This means uses natural language processing libraries such as NLTK and Transformers to analyze text and recognize emotions. For example, it recognizes the emotion "joy" from the text "I had fun today."

[0559] The server also includes a means for adjusting the response of the artificial intelligence based on the analyzed emotions, thereby enabling it to generate an appropriate response according to the user's emotions.

[0560] The server also contains a mechanism for storing the received answers in a database, which uses the Python pymysql library to connect to a MySQL database and store the answers.

[0561] The server also includes a means for selecting an optimal response based on emotion, the means including logic for selecting an optimal response from among the stored answers based on the analysis of the emotion engine.

[0562] Finally, the server has a means for returning a response to the user. This means includes logic for sending the selected response to the user's device. For example, if the user enters "I had fun today," the server will send the response "That's great! What did you enjoy?" to the user's device.

[0563] For example, if the user types "Today was a tough day," here's the prompt:

[0564] User: "Today was fun"

[0565] AI: "That's great! What was fun about it?"

[0566] User: "Today was tough."

[0567] AI: "That must have been tough. Maybe talking about it will make you feel a little better."

[0568] In this way, the server centrally manages responses from multiple AIs, creating a system that provides appropriate responses according to the user's emotions.

[0569] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0570] Step 1:

[0571] Accepting user input

[0572] A user inputs text into the system using a terminal. For example, the user inputs "I had a good time today." This input text is sent to the server.

[0573] Step 2:

[0574] Analyzing emotions with an emotion engine

[0575] The server receives the user's input text and passes it to the sentiment analysis engine. The sentiment analysis engine uses a natural language processing library (such as NLTK or Transformers) to analyze the text and recognize emotions. For example, it can recognize the emotion "joy" from the text "I had fun today." The analysis results are then returned to the server.

[0576] Step 3:

[0577] Send prompts to each AI

[0578] The server sends prompts to the generative AI model based on the analyzed emotions. For example, it sends a prompt such as "The user feels happy. Please ask him what he enjoyed." This prompt is then sent to each AI.

[0579] Step 4:

[0580] Receive answers from each AI

[0581] The server receives the answers from each AI through the API. For example, AI1 will respond, "That's great! What was fun about it?" These answers are then aggregated on the server.

[0582] Step 5:

[0583] Save the answers to the database

[0584] The server stores the received answers in a MySQL database. It uses the Python pymysql library to connect to the database and store the answers. For example, the answer "That's great! What did you enjoy?" is stored in the database.

[0585] Step 6:

[0586] Choose the best response based on emotion

[0587] The server selects the best response from the stored answers based on the analysis results of the emotion engine. For example, it selects the most appropriate response for the emotion of "joy." This selected response is stored on the server.

[0588] Step 7:

[0589] Returning a response to the user

[0590] The server sends the selected response to the user's device, where the user receives the response "That's great! What did you enjoy?" This response is displayed on the user's device.

[0591] (Application example 2)

[0592] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0593] Conventional AI systems have had difficulty providing appropriate responses and recommending content based on user emotions. In particular, there was no system that could centrally manage responses from multiple AIs and adjust responses based on user emotions. This could lead to inconsistent user experiences and reduced satisfaction.

[0594] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0595] In this invention, the server includes means for sending the same prompt content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending new prompts based on the received multiple answers, means for determining the credibility of the same answer content, means for supporting chat-type and image-generating AIs, means for analyzing emotions from user input text, means for adjusting the AI's response based on the analyzed emotions, and means for recommending content according to the emotions. This makes it possible to recommend appropriate responses and content according to the user's emotions.

[0596] "Multiple AIs" refers to multiple artificial intelligence systems that operate using different algorithms and datasets.

[0597] A "prompt" is an instruction or question given to an AI to help it perform a specific task.

[0598] An "answer" is the response or output that the AI ​​generates based on a prompt.

[0599] "Credibility" is a criterion for assessing whether answers from AI are accurate and trustworthy.

[0600] "Chat AI" refers to an artificial intelligence system designed for text-based dialogue.

[0601] "Image generation AI" refers to an artificial intelligence system designed to generate images.

[0602] "Means for analyzing emotions" refers to technology that identifies emotions from user input text and evaluates the type and intensity of those emotions.

[0603] "Means to adjust responses" refers to technology that changes the AI's responses based on analyzed emotions.

[0604] "Content recommendation means" refers to technology that suggests appropriate content such as movies, music, and articles based on the user's emotions and preferences.

[0605] An embodiment of this invention will now be described. First, the entire system is composed of a server, a user terminal, and multiple AIs. The server sends the same prompt content to multiple AIs and receives answers from them. It creates and sends a new prompt based on the multiple answers it receives. It also has the function of judging the credibility of the same answer content. Furthermore, it is compatible with chat-type and image-generation-type AIs.

[0606] The user terminal is a device such as a smartphone or PC, which provides an interface for users to input text. The text entered by the user is sent to the server, which analyzes the text for emotion. For emotion analysis, an emotion analysis library such as TextBlob is used. Based on the analyzed emotion, the server adjusts the AI's response and recommends appropriate content.

[0607] The specific hardware and software configuration is as follows: the server uses Flask (a Python web framework) and the TextBlob library for sentiment analysis. User devices are general devices such as smartphones and PCs, and communicate with the server via an internet connection.

[0608] For example, if a user types "I had fun today," the server receives the text and analyzes it for the emotion of "joy" using the TextBlob library. The server then recommends fun movies and upbeat music based on the emotion of "joy." An example of a specific prompt sentence is the text "I had fun today."

[0609] In this way, it becomes possible to provide appropriate responses and recommend content according to the user's emotions.

[0610] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0611] Step 1:

[0612] The user enters text into the terminal.

[0613] Input: Text entered by the user (e.g., "I had fun today")

[0614] Output: Text data sent from the terminal to the server

[0615] Step 2:

[0616] The server passes the received text to a sentiment analysis engine.

[0617] Input: Text data sent from the terminal

[0618] Output: Text data passed to the sentiment analysis engine

[0619] Step 3:

[0620] The server analyzes the sentiment of the text using a sentiment analysis engine.

[0621] Input: Text data passed to the sentiment analysis engine

[0622] Data Processing: Sentiment Analysis using the TextBlob Library

[0623] Output: Parsed emotion data (e.g., "joy")

[0624] Step 4:

[0625] The server sends prompts to multiple AIs based on the analyzed emotions.

[0626] Input: Parsed emotion data

[0627] Data processing: Generate prompts based on sentiment data

[0628] Output: Prompt data sent to multiple AIs

[0629] Step 5:

[0630] Multiple AIs generate answers based on the prompts and send them to the server.

[0631] Input: Prompt data sent from the server

[0632] Data calculations: Each AI generates an answer based on a prompt

[0633] Output: Answer data sent from the AI ​​to the server

[0634] Step 6:

[0635] The server receives answers from multiple AIs and judges their credibility.

[0636] Input: Answer data sent from AI

[0637] Data calculation: Evaluating the degree of agreement and reliability of answers

[0638] Output: Answer data with credibility assessed

[0639] Step 7:

[0640] The server recommends content that matches the user's emotions based on the most reliable answers.

[0641] Input: Answer data with credibility assessment

[0642] Data processing: Select content based on sentiment data and response data

[0643] Output: Recommended content data (e.g., "fun movies," "upbeat music")

[0644] Step 8:

[0645] The server transmits the recommended content to the user terminal.

[0646] Input: Recommended content data

[0647] Output: Content data sent to the user device

[0648] Step 9:

[0649] The user terminal displays the received content to the user.

[0650] Input: Content data sent from the server

[0651] Output: The content displayed to the user (e.g., "Fun movie," "Upbeat music")

[0652] Example 3

[0653] Next, a description will be given of a third embodiment of the third embodiment. 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."

[0654] In conventional AI systems, responses to user input often lacked credibility and specificity, and there were limited ways to effectively utilize responses from multiple AIs. Furthermore, there was a lack of interfaces that supported both chat AI and image generation AI, limiting the quality and variety of information available to users.

[0655] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0656] In this invention, the server includes means for receiving user input, means for sending the received input as a prompt to the generative AI model, means for receiving answers from the generative AI model, means for analyzing answers from the generative AI models, means for extracting keywords from the analyzed answers and generating a new prompt, means for resending the new prompt to the generative AI model, means for evaluating the credibility of answers from the generative AI models, and means for displaying answers from the chat AI and the image generation AI. This allows users to obtain highly reliable information and further enables them to effectively utilize both text information and image information.

[0657] The "means for receiving user input" is an interface through which the user inputs questions or requests to the system.

[0658] "Means for sending received input as a prompt to the generative AI model" is a function for sending input from the user to the generative AI model in an appropriate format.

[0659] "Means for receiving an answer from a generative AI model" is a function for receiving an answer returned from a generative AI model.

[0660] "Means for analyzing answers from multiple generative AI models" refers to a function for comparing and analyzing answers obtained from multiple generative AI models.

[0661] "Means of extracting keywords from analyzed answers and generating new prompts" is a function that extracts important keywords from the analysis results and creates new prompts based on them.

[0662] The "means for resending a new prompt to the generation AI model" is a function for resending the generated new prompt to the generation AI model.

[0663] "Means for evaluating the credibility of answers from multiple generative AI models" is a function for evaluating the reliability of answers obtained from multiple generative AI models.

[0664] "Means for displaying answers from chat AI and image generation AI" is a function for appropriately displaying text answers from chat AI and image answers from image generation AI to the user.

[0665] MODE FOR CARRYING OUT THE INVENTION

[0666] This invention is a system that receives user input, sends prompts to a generative AI model based on that input, analyzes the responses from the generative AI model, and generates new prompts. It also has the function of evaluating the credibility of responses from multiple generative AI models and appropriately displaying the responses of the chat AI and image generation AI.

[0667] Hardware and software used

[0668] The system uses the following hardware and software:

[0669] Server: A central processing unit for receiving user input, sending prompts to the generative AI model, and analyzing the answers.

[0670] Terminal: A device used by a user to input information (e.g., a computer or smartphone)

[0671] Generative AI model: An artificial intelligence model (e.g., GPT-4) for generating answers to prompts.

[0672] Chat AI: Artificial intelligence that generates text-based responses (e.g., ChatGPT)

[0673] Image generation AI: Artificial intelligence that generates answers in image format (e.g., DALL-E)

[0674] Data processing and calculation

[0675] The server receives input from the user and sends it as a prompt to the generative AI model. After receiving the answer from the generative AI model, the server analyzes the answers from multiple generative AI models and extracts keywords. It generates a new prompt based on the extracted keywords and sends it again to the generative AI model. It then compares the answers from multiple generative AI models and evaluates their credibility. Finally, the answers from the chat AI and image generation AI are displayed appropriately on the user's device.

[0676] Specific examples

[0677] For example, if a user types "Please tell me your ideas for a new project" on their device, the server sends this input as a prompt to the generative AI model. The generative AI model responds with a response such as "One option is developing environmentally friendly products." The server analyzes this response and extracts the keywords "environmentally friendly products" and "development." The server then generates a new prompt, "Please tell me specific methods for developing environmentally friendly products," and sends it again to the generative AI model. If the generative AI model responds with a response such as "Using recycled materials is effective," the server compares the responses from multiple generative AI models and evaluates their credibility. Finally, the server displays the text response from the chat AI and the image response from the image generation AI on the user's device.

[0678] Prompt Sentence Examples

[0679] "Tell me your ideas for a new project."

[0680] "Please tell me the specific methods for developing environmentally friendly products."

[0681] "What factors contributed to the success of this project?"

[0682] "What are the risks of the project?"

[0683] In this way, the user can obtain specific and highly reliable information through the system. The flow of the identification process in the third embodiment will be described with reference to FIG.

[0684] Step 1:

[0685] The user types at the terminal.

[0686] Specific Action: The user uses the device interface to enter a question or request, for example, "Can you give me some ideas for a new project?"

[0687] Input: User question or request

[0688] Output: User input data

[0689] Step 2:

[0690] The server sends the user's input as a prompt to the generative AI model.

[0691] Specific operation: The server receives the user's input and sends it to the generative AI model as a prompt sentence.

[0692] Input: User-entered data

[0693] Output: The prompt sent to the generative AI model

[0694] Step 3:

[0695] The generative AI model generates answers to the prompts and returns them to the server.

[0696] How it works: The generative AI model generates an answer based on the prompt it receives and returns the answer to the server, such as "We can consider developing environmentally friendly products."

[0697] Input: A prompt sent to the generative AI model

[0698] Output: Answer from the generative AI model

[0699] Step 4:

[0700] The server analyzes the answers from multiple generative AI models.

[0701] How it works: The server receives answers from multiple generative AI models and compares and analyzes them. During analysis, the content of the answers is compared to find similarities and differences.

[0702] Input: Answers from multiple generative AI models

[0703] Output: Analysis results (commonalities and differences)

[0704] Step 5:

[0705] The server extracts keywords from the analyzed answers and generates a new prompt.

[0706] Specific operation: The server extracts important keywords from the analysis results and creates a new prompt based on them. For example, the keywords extracted are "environmentally friendly products" and "development."

[0707] Input: Analysis results

[0708] Output: New prompt text

[0709] Step 6:

[0710] The server resends a new prompt to the generative AI model.

[0711] Specific operation: The server sends the newly generated prompt to the generation AI model again, for example, "Please tell me specific ways to develop environmentally friendly products."

[0712] Input: New prompt text

[0713] Output: The prompt sentence resubmitted to the generative AI model

[0714] Step 7:

[0715] The generative AI model generates answers to new prompts and returns them to the server.

[0716] How it works: The generative AI model generates an answer based on the new prompt and returns it to the server, for example, "Using recycled materials is effective."

[0717] Input: The prompt sentence resubmitted to the generative AI model

[0718] Output: A new answer from the generative AI model

[0719] Step 8:

[0720] The server evaluates the credibility of answers from multiple generative AI models.

[0721] Specific operation: The server compares answers from multiple generative AI models, and if there are answers with the same content, it rates their credibility higher.

[0722] Input: Answers from multiple generative AI models

[0723] Output: Credibility evaluation result

[0724] Step 9:

[0725] The server displays the answers from the chat AI and image generation AI.

[0726] Specific operation: The server receives both the text response from the chat AI and the image response from the image generation AI, and displays them appropriately on the user's device. For example, the text response might say "Use recycled materials," and the image response might show an image of a product made from recycled materials.

[0727] Input: Credibility assessment results, text response from chat AI, image response from image generation AI

[0728] Output: Text and image answers displayed on the user's device

[0729] (Application example 3)

[0730] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[0731] In modern brick-and-mortar stores, it is difficult for users to quickly and accurately obtain detailed information about products. There is also a need to provide a more natural and satisfying shopping experience by providing responses that reflect the user's emotions. Furthermore, there is a lack of a way to integrate information from multiple information processing devices and provide highly reliable information.

[0732] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0733] In this invention, the server includes means for sending the same instruction content to a plurality of information processing devices, means for receiving responses from the plurality of information processing devices, means for creating and sending new instructions based on the received plurality of responses, means for determining the credibility of the same response content, means for supporting both interactive information processing devices and image generation information processing devices, means for recognizing a user's emotions and generating a response accordingly, means for generating a new question in response to a user's question based on responses from the plurality of information processing devices and providing detailed information, means for analyzing the user's facial expression and tone of voice, means for analyzing information from the plurality of information processing devices and providing highly credible information when the user scans an item, and means for generating images of clothes the user wants to try on and providing a try-on simulation. This allows users to quickly and accurately obtain information in a physical store and receive responses according to their emotions, enabling a more natural and satisfying shopping experience.

[0734] An "information processing device" is a device for processing and analyzing data.

[0735] The "instruction content" is the content that specifies the processing or operation that should be performed by the information processing device.

[0736] A "response" is information that an information processing device returns based on the contents of an instruction.

[0737] "Credibility" is the degree to which information or responses are accurate and trustworthy.

[0738] An "interactive information processing device" is a device that processes information through interaction with a user.

[0739] An "image generation type information processing device" is an information processing device for generating images.

[0740] The "means for recognizing emotions" refers to a method or device for analyzing and recognizing a user's emotions.

[0741] A "means for generating new questions" is a method or device for creating new questions based on the responses received.

[0742] "Detailed information" is specific and in-depth information about a particular subject.

[0743] The "means for analyzing facial expressions and tone of voice" refers to a method or device for analyzing a user's facial expressions and tone of voice.

[0744] "Means for scanning products" refers to a method or device for reading product information.

[0745] "Try-on simulation" involves generating images of the clothes the user wants to try on, providing the experience of virtually trying them on.

[0746] A system for implementing this invention has the functions of sending the same instruction content to multiple information processing devices, receiving responses from the multiple information processing devices, and creating and sending new instructions based on the responses.It also has the functions of determining the authenticity of the same response content and supporting both interactive information processing devices and image generation information processing devices.It also includes the functions of recognizing the user's emotions and generating responses accordingly.

[0747] Hardware and software used

[0748] Hardware: Camera (to capture the user's facial expressions), smartphone (as a user interface)

[0749] software:

[0750] OpenAI API: To get answers from AI

[0751] EmotionRecognizer: To recognize user emotions

[0752] OpenCV: for processing camera images

[0753] System Operation

[0754] The server first captures the user's facial expression using a camera and recognizes their emotions using EmotionRecognizer. Next, when the user enters a question about a product via their smartphone, the server sends the question to multiple information processing devices using the OpenAI API. After receiving responses from the information processing devices, the server analyzes the responses and selects the most reliable information.

[0755] The server then extracts keywords from the response and generates a new prompt. Based on this new prompt, the server sends a question to the information processing device to obtain more detailed information. The obtained information is then displayed on the smartphone in a format that reflects the user's emotions.

[0756] Specific examples

[0757] For example, if a user asks, "I'd like to know more about this product," the server generates the following prompt and sends it to the information processing device:

[0758] Example prompt sentence:

[0759] A user has asked, "I want to know more about this product." Please provide the following information:

[0760] Product details

[0761] Other users' reviews

[0762] How to use

[0763] Related products

[0764] After receiving the responses from the information processing devices, the server analyzes the responses and selects information with high credibility. For example, if the server receives responses such as "This product is of high quality" from multiple information processing devices, it will rate the credibility of that information as high.

[0765] Furthermore, when a user generates an image of the clothes they want to try on, the server uses an image generation information processing device to provide a try-on simulation, allowing the user to virtually try on the clothes without actually trying them on.

[0766] In this way, the system provides users with fast and accurate information and generates emotionally appropriate responses, resulting in a more natural and satisfying shopping experience.

[0767] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0768] Step 1:

[0769] The server captures the user's facial expressions through a camera. The input is the camera image, and the output is the captured image data. Specifically, the server acquires the image from the camera in real time, divides the image into frames, and saves them as image data.

[0770] Step 2:

[0771] The server uses EmotionRecognizer to recognize the user's emotions from the captured image data. The input is image data, and the output is recognized emotion data. Specifically, the server inputs the image data into EmotionRecognizer and applies the emotion recognition algorithm to analyze the user's emotions.

[0772] Step 3:

[0773] A user inputs a question about a product through a smartphone. The input is the user's question text, and the output is the question data sent to the server. In concrete terms, the user inputs a question using a smartphone application, and the question is sent to the server.

[0774] Step 4:

[0775] The server uses the OpenAI API to send user questions to multiple information processing devices. The input is the question data, and the output is the instructions sent to the information processing devices. Specifically, the server inputs the question data into the OpenAI API and sends the same instructions to multiple information processing devices.

[0776] Step 5:

[0777] The server receives responses from multiple information processing devices. The input is the response data from the information processing devices, and the output is the received response data. Specifically, the server collects responses from each information processing device through the OpenAI API and saves them as a single dataset.

[0778] Step 6:

[0779] The server analyzes the received response data and selects the most credible information. The input is the response data, and the output is the most credible information. Specifically, the server inputs the response data into an analysis algorithm, and if there are multiple responses with the same content, it rates the most credible one.

[0780] Step 7:

[0781] The server extracts keywords from the received response and generates a new prompt. The input is credible information, and the output is a new prompt. Specifically, the server extracts important keywords from the credible information and creates a new prompt based on them.

[0782] Step 8:

[0783] The server then sends a new question to the information processing device based on the new prompt and obtains detailed information. The input is the new prompt text, and the output is the detailed information. Specifically, the server inputs the new prompt text into the OpenAI API and obtains detailed information from the information processing device.

[0784] Step 9:

[0785] The server then displays the acquired detailed information on the smartphone in a format that corresponds to the user's emotions. The input is detailed information and emotional data, and the output is a response that is displayed on the smartphone. Specifically, the server combines the detailed information and emotional data to generate a response that is optimal for the user and sends it to the smartphone.

[0786] Step 10:

[0787] The user checks the information displayed on their smartphone and asks additional questions or takes action as necessary. The input is the response displayed on the smartphone, and the output is the user's next action. Specifically, the user looks at the smartphone screen to check the information, and if necessary, enters additional questions or purchases a product.

[0788] 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.

[0789] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by 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.

[0790] Another example of generative AI is Gemini (registered trademark) (Internet search engine). <url: https: gemini.google.com ?hl="ja">) are mentioned.

[0791] 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.

[0792] [Second embodiment]

[0793] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0794] 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.

[0795] 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).

[0796] 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.

[0797] 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.

[0798] 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).

[0799] 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.

[0800] 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.

[0801] 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.

[0802] 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.

[0803] In the smart glasses 214, 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.

[0804] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0805] "Example 1"

[0806] The system of the present invention has a function to send the same prompt content to multiple AIs via API. Specifically, the system sends the same prompt to AIs such as OpenAI via Azure and Bard via GCP.

[0807] "Example 2"

[0808] In addition, as a means of receiving answers from multiple AIs, it has the function of receiving answers from each AI via API and managing them centrally. Specifically, it receives answers from each AI and stores them in a database.

[0809] "Example 3"

[0810] Furthermore, as a means of creating and sending new prompts based on multiple received answers, the system has the ability to analyze the received answers, generate new prompts based on them, and send them to the AI. Specifically, it extracts keywords from the received answers and generates new prompts based on them.

[0811] "Example 4"

[0812] In addition, as a means of determining the credibility of the same answer, the system has a function to compare answers received from multiple AIs and, if there are answers with the same content, rate the credibility of that answer highly. Specifically, the system compares answers received from multiple AIs and, if there are answers with the same content, rates the credibility of that answer highly.

[0813] "Example 5"

[0814] Furthermore, as a means of supporting both chat and image generation AI, it has an interface that supports both chat and image generation AI. Specifically, it has the function of receiving both text responses from chat AI and image responses from image generation AI and displaying them appropriately.

[0815] The processing flow of each embodiment will be described below.

[0816] "Example 1"

[0817] Step 1: User enters prompt.

[0818] Step 2: The system sends identical prompts to each AI via API (e.g., OpenAI via Azure, Bard via GCP).

[0819] Step 3: The system receives the answers from each AI via API.

[0820] Step 4: The system centralizes the responses received and stores them in a database.

[0821] "Example 2"

[0822] Step 1: The system retrieves the answers from each AI stored in the database.

[0823] Step 2: The system analyzes the answers obtained and extracts keywords.

[0824] Step 3: The system generates a new prompt based on the extracted keywords.

[0825] Step 4: Send the new system-generated prompts to each AI.

[0826] "Example 3"

[0827] Step 1: The system compares the answers it receives from multiple AIs.

[0828] Step 2: If there is an answer with the same content, the system will rate the credibility of that answer highly.

[0829] Step 1: The system receives both a text response from the chat AI and an image response from the image generation AI.

[0830] Step 2: The system displays the received text and image answers appropriately.

[0831] Example 1

[0832] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0833] In conventional systems, it was difficult to send the same prompt to multiple generative AI models and compare their responses. Furthermore, there was a lack of means to determine the credibility of the responses from each generative AI model, which sometimes raised doubts about the reliability of the information users received. Furthermore, there was no general-purpose system that could handle chat-based and image-based generative AI models.

[0834] 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.

[0835] In this invention, the server includes a means for a user to input a prompt text, a means for a terminal to send the prompt text to the server, a means for the server to receive the prompt text, a means for the server to send the prompt text to a generative AI model via a cloud service, a means for the server to receive responses from each generative AI model, a means for the server to return the response to the user, a means for sending the same prompt content to multiple generative AI models, a means for receiving answers from multiple generative AI models, a means for creating and sending a new prompt based on the multiple received answers, a means for determining the credibility of the same answer content, and a means for supporting chat-based and image-based generative AI models. This makes it possible to compare responses from multiple generative AI models and determine their credibility. Furthermore, a general-purpose system compatible with chat-based and image-based generative AI models can be provided.

[0836] "User" means an individual or entity that utilizes the system to enter prompts and receive responses from the generative AI model.

[0837] A "terminal" is a device on which a user inputs a prompt and sends it to a server, and includes a PC, smartphone, tablet, etc.

[0838] The "server" is a computer system whose role is to receive prompts, send them to the generative AI model, receive responses, and return them to the user.

[0839] A "prompt" is a question or instruction text that a user enters into a generative AI model.

[0840] "Cloud services" is a general term for computing resources and services provided over the internet, and is a platform that enables access to generative AI models.

[0841] A "generative AI model" is an artificial intelligence algorithm or system that generates a response to a prompt from a user.

[0842] A "response" is an answer or piece of information that a generative AI model generates based on a prompt.

[0843] "Credibility" is a criterion for assessing whether the responses from a generative AI model are accurate and trustworthy.

[0844] A "chat-based generative AI model" is a generative AI model designed for text-based dialogue.

[0845] An "image generation system generative AI model" is a generative AI model designed to generate images based on prompt text.

[0846] A "new prompt" is a new question or instruction text that is created based on multiple responses received.

[0847] This invention is a system that sends the same prompt sentence to multiple generative AI models and compares their responses. A specific embodiment of this system is described below.

[0848] The user inputs a prompt using a terminal, which can be a PC, smartphone, tablet, or other device, and sends the input prompt to the server. The server receives the prompt and sends it to the generative AI model via a cloud service.

[0849] The server uses a platform commonly used for cloud services, such as Azure or GCP, to access the generative AI model. Specifically, it uses the Azure SDK or GCP SDK to generate an API request and includes the prompt statement as the payload.

[0850] The server receives responses from each generative AI model and returns them to the user. Responses are answers or information generated by the generative AI model based on the prompt. The server compiles the received responses and returns them to the device, which then displays them to the user.

[0851] As a specific example, if a user enters a prompt such as "What's the weather like today?", the server sends this prompt to multiple generative AI models via a cloud service. For example, the server sends the same prompt to a generative AI model provided via Azure and a generative AI model provided via GCP. After receiving responses from each generative AI model, the server returns these responses to the user. The user will see the message "Generative AI model A: Today's weather is sunny. Generative AI model B: Today's weather is cloudy" displayed on their device screen.

[0852] This system allows users to compare responses from multiple generative AI models and judge their credibility. It also provides a general-purpose system that supports chat and image generation generative AI models.

[0853] Examples of prompts include:

[0854] "What's the weather like today?"

[0855] "Please tell me the latest news."

[0856] How do I write a Hello World program in Python?

[0857] In this way, users can ask a variety of questions and give instructions to the generative AI model and use the responses it receives.

[0858] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0859] Step 1:

[0860] The user enters a prompt statement.

[0861] Specifically, the user enters "What's the weather like today?" into the input field of the terminal. The entered prompt text is temporarily stored in the terminal's memory. Input: The prompt text entered by the user. Output: The prompt text stored in the terminal.

[0862] Step 2:

[0863] The terminal sends the prompt to the server.

[0864] The device uses the JavaScript fetch API to send an HTTP POST request to the server containing the prompt text. The request also includes the user's identity and a timestamp. Input: The prompt text stored on the device. Output: The HTTP request sent to the server.

[0865] Step 3:

[0866] The server receives the prompt.

[0867] The server uses the Node.js Express framework to extract the prompt text from the received HTTP request. The extracted prompt text is stored in the server's memory. Input: HTTP request sent to the server. Output: Prompt text stored on the server.

[0868] Step 4:

[0869] The server sends the prompt to the generative AI model via a cloud service.

[0870] The server uses the Azure SDK to send a prompt to the API of Generative AI Model A. At the same time, it uses the GCP SDK to send the same prompt to the API of Generative AI Model B. Input: Prompt text saved on the server. Output: Prompt text sent to Generative AI Model A and Generative AI Model B.

[0871] Step 5:

[0872] The server receives the response from each generative AI model.

[0873] The server waits for HTTP responses from Generative AI Model A and Generative AI Model B, and receives each response. The received responses are stored in the server's memory. Input: HTTP responses from Generative AI Model A and Generative AI Model B. Output: Responses stored on the server.

[0874] Step 6:

[0875] The server returns the response to the user.

[0876] The server compiles the received responses in JSON format and returns them to the device. The device then displays this to the user. Specifically, the browser screen displays "Generated AI model A: Today's weather is sunny. Generated AI model B: Today's weather is cloudy." Input: Response stored on the server. Output: Response displayed to the user.

[0877] (Application example 1)

[0878] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0879] Conventional systems using generative AI models lacked a means to efficiently compare and integrate responses from multiple AIs, making it difficult to provide information to users in an optimal format. Furthermore, the lack of consideration for implementation as an application installed on a specific device resulted in poor user convenience. This posed a challenge, limiting the scope of use of generative AI models.

[0880] 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.

[0881] In this invention, the server includes means for sending the same prompt content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending new prompts based on the received multiple answers, means for determining the credibility of the same answer content, means for supporting chat-type and image generation-type AIs, means for comparing and integrating the received multiple answers and providing them to the user in an optimal form, and means including an application to be installed on a smartphone. This makes it possible to efficiently compare and integrate answers from multiple generative AI models and provide information to the user in an optimal form.

[0882] "Artificial intelligence" refers to computer systems that have the ability to analyze data, learn, and make decisions.

[0883] A "prompt" is an input sentence that causes an artificial intelligence to perform a specific task.

[0884] An "answer" is the output generated by an artificial intelligence in response to a prompt.

[0885] "Credibility" is a criterion for evaluating the accuracy and reliability of an answer.

[0886] "Chat AI" is AI for conducting text-based conversations.

[0887] "Image generation artificial intelligence" is artificial intelligence that has the ability to generate images.

[0888] "Comparison and synthesis" is the process of evaluating multiple responses and combining them into one optimal answer.

[0889] A "smartphone" is a mobile phone that can connect to the Internet and run applications.

[0890] An "application" is a software program designed to perform a specific task.

[0891] A system for implementing this invention sends the same prompt content to multiple artificial intelligences, receives responses from the multiple artificial intelligences, compares and integrates them, and provides them to a user. The system is implemented as an application installed on a smartphone.

[0892] Hardware and software used

[0893] Hardware: Smartphone

[0894] Software: Python, cloud service APIs (Azure's OpenAI API, GCP's Bard API)

[0895] System Operation

[0896] 1. Entering a prompt: The user enters a prompt through a smartphone application, for example, "Tell me about the city of the future."

[0897] 2. Sending prompts: The server sends the input prompt to OpenAI via Azure and Bard via GCP, allowing the same prompt to be sent to multiple AIs.

[0898] 3. Receiving answers: The server receives answers from each AI. For example, OpenAI answers "The city of the future is a place where smart technology and sustainable energy solutions come together," while Bard answers "The city of the future is an environment where AI and IoT are highly integrated."

[0899] 4. Comparing and integrating answers: The server compares the multiple answers received and integrates them in the most optimal way. For example, it might integrate them as "The city of the future is an environment where smart technology and sustainable energy solutions are combined, and AI and IoT are highly integrated."

[0900] 5. Providing to users: The integrated answers are provided to users through a smartphone application, allowing them to obtain information from multiple AIs at once.

[0901] Specific examples

[0902] If the user types the prompt "Tell me about the city of the future," the system will:

[0903] Example prompt: Tell me about the city of the future.

[0904] Based on this prompt, OpenAI via Azure and Bard via GCP provide their respective answers, which are then integrated and presented to the user. For example, the integrated answer looks like this:

[0905] Example of an integrated answer: The city of the future is a blend of smart technology and sustainable energy solutions, with highly integrated AI and IoT.

[0906] In this way, users can efficiently obtain information from multiple generative AI models and utilize it in the most optimal way.

[0907] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0908] Step 1:

[0909] A user starts an application on their smartphone and inputs a prompt. The input prompt is text data such as "Tell me about the city of the future." The input prompt is then sent to the server.

[0910] Step 2:

[0911] The server sends the received prompt sentence to OpenAI via Azure and Bard via GCP. Specifically, it sends the prompt sentence using the API of each cloud service. The input is the prompt sentence, and the output is the answer from each AI.

[0912] Step 3:

[0913] The server receives answers from each AI. For example, OpenAI returns text data such as "The city of the future is a place where smart technology and sustainable energy solutions are combined," and Bard returns text data such as "The city of the future is an environment where AI and IoT are highly integrated." The input is the answer from each AI, and the output is a list of the received answers.

[0914] Step 4:

[0915] The server compares and integrates the multiple responses it receives. Specifically, it analyzes the content of each response and integrates them while eliminating duplicates. For example, it might integrate the following statement: "The city of the future is an environment where smart technology and sustainable energy solutions are combined, and AI and IoT are highly integrated." The input is a list of received responses, and the output is the integrated response.

[0916] Step 5:

[0917] The server sends the integrated answer to a smartphone application, whose input is the integrated answer and whose output is text data displayed on the user's smartphone.

[0918] Step 6:

[0919] The user can view the integrated answer through a smartphone application, which allows the user to obtain information from multiple generative AI models at once. The input is the integrated answer sent from the server, and the output is text data that the user can view.

[0920] Example 2

[0921] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0922] In conventional AI systems, it was difficult to centrally manage responses from multiple AIs and determine their credibility. Furthermore, the system lacked the ability to efficiently store received responses and search and view them later. This made it time-consuming for users to compare and consider responses from multiple AIs, hindering efficient use.

[0923] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0924] In this invention, the server includes means for sending the content of the same input sentence to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending a new input sentence based on the multiple received answers, means for determining the credibility of the same answer content, means for handling chat-type and image-generating AIs, means for centrally managing answers from each AI, means for saving received answers in a database, and means for searching and viewing saved answers. This allows answers from multiple AIs to be efficiently managed, their credibility to be determined, and users to easily search and view them.

[0925] "Artificial intelligence" refers to computer systems that have the ability to learn from data and perform specific tasks automatically.

[0926] An "input sentence" is a textual instruction or question given by a user to an artificial intelligence.

[0927] An "answer" is a text or other form of output that an artificial intelligence generates based on an input sentence.

[0928] "Credibility" is a criterion for assessing whether answers from artificial intelligence are accurate and trustworthy.

[0929] "Chat AI" is AI designed for text-based conversations.

[0930] "Image generation artificial intelligence" is artificial intelligence that has the ability to generate images based on input text and other data.

[0931] "Centralized management" means integrating and managing multiple pieces of information and data in one system or location.

[0932] A "database" is a system for efficiently storing, searching, and managing data.

[0933] "Searching and viewing" means searching for stored data based on specific conditions and checking its contents.

[0934] The present invention relates to a system for centrally managing responses from multiple artificial intelligences and determining their credibility. Specific embodiments of this system will be described below.

[0935] The server sends the prompt text entered by the user to multiple generative AI models. Generative AI models include models specialized in text generation and image generation. Specific examples of generative AI models include OpenAI's GPT-3 and Google's BERT.

[0936] The server receives the prompt text entered by the user and sends an API request to each generative AI model. The API request is made using an HTTP POST request, sending data including the prompt text. The answer from each generative AI model is returned in JSON format.

[0937] The server receives the answers from each generative AI model and stores them in a database, typically a relational database such as MySQL or PostgreSQL. The stored data includes prompt statements and the corresponding answers from each generative AI model.

[0938] Users can search and view past prompts and their responses through the system interface. For example, they can display a list of responses from each generative AI model to the prompt, "Tell me about the future of AI-based text generation."

[0939] For example, if a user enters the prompt "Please tell me about the future of AI text generation," the server sends this prompt to each generative AI model. The server receives the answers from each generative AI model and stores them in a database. The user can later compare the answers of each generative AI model to this prompt.

[0940] Examples of prompts include:

[0941] "Please tell me about the future of AI text generation."

[0942] "Please explain the effects of climate change."

[0943] "Tell me about the latest technology trends"

[0944] This system efficiently manages answers from multiple generative AI models, determines their credibility, and allows users to easily search and view them.

[0945] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0946] Step 1:

[0947] The user enters a prompt statement.

[0948] The user enters a prompt through the system interface, for example, by typing "Tell me about the future of AI-generated text" into a text box on a web browser and clicking the submit button.

[0949] Input: The prompt text entered by the user.

[0950] Output: The prompt text sent to the server.

[0951] Step 2:

[0952] The server sends the prompt sentence to each generative AI model.

[0953] The server receives the prompt sentence entered by the user and sends an API request to each generative AI model. Specifically, it sends the prompt sentence to the API endpoint of the generative AI model using an HTTP POST request.

[0954] Input: The prompt received from the user.

[0955] Output: API requests sent to each generative AI model.

[0956] Step 3:

[0957] The server receives the answers from each generative AI model.

[0958] The server receives API responses from each generative AI model. Responses are usually returned in JSON format. For example, from GPT-3, the server receives the answer, "The future of AI text generation is very bright."

[0959] Input: API response from each generative AI model.

[0960] Output: The response data received by the server.

[0961] Step 4:

[0962] The server stores the received response in a database.

[0963] The server saves the received answers in a database. Specifically, it uses a relational database such as MySQL or PostgreSQL to store the prompt sentences and the corresponding answers of each generative AI model in a table.

[0964] Input: The answer data received from each generative AI model.

[0965] Output: Response data stored in a database.

[0966] Step 5:

[0967] Users can search and view saved answers.

[0968] Users can search and view past prompts and their responses through the system interface. For example, the system displays a list of responses from each generative AI model to the prompt, "Tell me about the future of AI-based text generation."

[0969] Input: The user's search query.

[0970] Output: The prompt and answer data displayed as search results.

[0971] (Application example 2)

[0972] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0973] Conventional AI systems have difficulty managing responses from multiple AIs in a unified manner and providing users with the most appropriate answer. Furthermore, they lacked the means to determine the credibility of the answers, making it impossible to provide users with highly reliable information. Furthermore, they lacked the ability to save answers in a database and refer to them later, making efficient information management impossible.

[0974] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0975] In this invention, the server includes means for sending the same input content to multiple AIs, means for receiving answers from the multiple AIs, means for creating and sending new input content based on the received multiple answers, means for determining the credibility of the same answer content, means for handling chat-type and image-generating AIs, means for saving answers from the multiple AIs in a database, means for selecting the optimal answer from the saved answers, and means for providing the selected optimal answer to the user. This makes it possible to centrally manage answers from the multiple AIs and provide the user with a highly reliable optimal answer.

[0976] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and learn, reason, and self-correct.

[0977] "Input" refers to questions or instructions provided by users to the artificial intelligence.

[0978] An "answer" is a response or information generated by artificial intelligence based on input.

[0979] A "database" is a system for efficiently storing, managing, and retrieving information.

[0980] "Credibility" is an indicator of whether information or answers are accurate and trustworthy.

[0981] "Chat AI" is AI that can conduct text-based conversations.

[0982] "Image generation artificial intelligence" is artificial intelligence that can generate images based on input content.

[0983] The "best answer" is the most appropriate and reliable answer selected from multiple answers.

[0984] "User" refers to a person or organization that uses the system to ask questions or give instructions.

[0985] "Cloud services" refer to computer resources and services provided over the Internet.

[0986] A "network" is a system in which multiple computers and devices are connected to each other and used to exchange information.

[0987] The system for carrying out this invention has the function of sending the same input content to multiple AIs, receiving responses from the multiple AIs, and managing them in a unified manner. A specific embodiment of this system will be described below.

[0988] System Configuration

[0989] The server includes the following means:

[0990] 1. Input content transmission means: Sends questions and instructions (prompt text) entered by the user to multiple artificial intelligences.

[0991] 2. Answer receiving means: Receive answers from multiple artificial intelligences.

[0992] 3. Method for creating new input content: Create new input content based on the multiple answers received and send it to the artificial intelligence again.

[0993] 4. Credibility judgment method: judge the credibility of the same answers.

[0994] 5. Database storage means: Responses from multiple artificial intelligences are stored in a database.

[0995] 6. Best answer selection method: Select the best answer from the saved answers.

[0996] 7. User-provided means: The selected best answer is provided to the user.

[0997] Hardware and software used

[0998] Hardware: Servers and user devices (smartphones, PCs, etc.) that can connect to the Internet.

[0999] Software: Python, SQLite, APIs for various artificial intelligence models (e.g., GPT-4, BERT, T5, etc.).

[1000] Data processing and calculation

[1001] The server receives input from the user and sends it to multiple AIs. After receiving answers from each AI, the server stores these answers in a database. The server selects the best answer from the stored answers and provides it to the user. The credibility judgment means evaluates the credibility of the same answer content and provides highly reliable information.

[1002] Specific examples

[1003] When a user asks, "What's the weather like today?", the server sends this prompt to multiple AIs. For example, GPT-4 may give an answer like, "It's sunny today," BERT may give an answer like, "It's cloudy today," and T5 may give an answer like, "It's going to rain today." These answers are stored in a database and evaluated using a credibility assessment tool. The best answer is selected and served to the user.

[1004] Example prompt sentence:

[1005] What's the weather like today?

[1006] In this way, it becomes possible to centrally manage answers from multiple AIs and provide users with the most reliable and optimal answers.

[1007] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1008] Step 1:

[1009] The user uses the terminal to enter input (prompt text).

[1010] Input: The prompt text entered by the user (e.g., "What's the weather like today?").

[1011] Output: The input is sent to the server.

[1012] Specific operation: When the user enters a question in the input field of the terminal and presses the send button, the prompt sentence is sent to the server.

[1013] Step 2:

[1014] The server receives the input and sends it to multiple generative AI models.

[1015] Input: The prompt text sent by the user.

[1016] Output: The prompt sentence is sent to multiple generative AI models.

[1017] Specific operation: The server sends the received prompt sentence to the API endpoint of each generative AI model (GPT-4, BERT, T5, etc.).

[1018] Step 3:

[1019] The server receives the answers from each generative AI model.

[1020] Input: The prompt sent to each generative AI model.

[1021] Output: The answer from each generative AI model.

[1022] Specific operation: Each generative AI model generates an answer based on the prompt sentence and sends the answer back to the server.

[1023] Step 4:

[1024] The server stores the received response in a database.

[1025] Input: The answers from each generative AI model.

[1026] Output: The answer is saved in the database.

[1027] Specific operation: The server stores the received answer in an SQLite database, along with information to identify which generative AI model the answer came from.

[1028] Step 5:

[1029] The server selects the best answer from the stored answers.

[1030] Input: Multiple answers stored in a database.

[1031] Output: The best answer.

[1032] What it does: The server analyzes the stored answers and runs an algorithm to select the best answer based on the authenticity and content of the answer.

[1033] Step 6:

[1034] The server provides the best answer to the user.

[1035] Input: Best answer.

[1036] Output: The answer provided to the user.

[1037] Specific operation: The server sends the selected optimal answer to the user's device and displays it on the user's screen.

[1038] In this way, it is possible to centrally manage responses from multiple generative AI models to prompts entered by users, and provide highly reliable and optimal answers.

[1039] Example 3

[1040] Next, a description will be given of Example 3 of Form Example 3. 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."

[1041] Conventional systems have difficulty effectively analyzing responses from multiple generative AI models and generating new prompts. Furthermore, they lack a means to determine the credibility of responses from multiple generative AI models, making it impossible to provide users with reliable information. Furthermore, there was no unified interface for chat and image generation generative AI models, resulting in an inconsistent user experience.

[1042] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1043] In this invention, the server includes means for receiving input from a user, means for sending prompts to multiple generative AI models, means for receiving responses from the multiple generative AI models, means for analyzing the received responses and extracting keywords, means for generating new prompts based on the extracted keywords, means for comparing responses from the multiple generative AI models and determining their credibility, and means for supporting chat-based and image-based generative AI models. This makes it possible to effectively analyze responses from multiple generative AI models based on user input and generate new prompts. Furthermore, by determining the credibility of responses from multiple generative AI models, highly reliable information can be provided to the user. Furthermore, by providing a unified interface for chat-based and image-based generative AI models, a consistent user experience can be achieved.

[1044] A "user" is an entity that inputs prompts into the system and receives answers from the generative AI model.

[1045] A "server" is a device or system that receives input from users, sends prompts to multiple generative AI models, and receives and analyzes responses.

[1046] A "generative AI model" is an artificial intelligence model that generates answers such as text and images based on a given prompt.

[1047] A "prompt" is text that contains instructions or questions that are input to a generative AI model.

[1048] An "answer" is an output result, such as text or an image, that a generative AI model generates based on a prompt.

[1049] "Keywords" are important words or phrases extracted by analyzing the answers from the generative AI model.

[1050] "Credibility" is the reliability of an answer, assessed based on the degree to which answers from multiple generative AI models match.

[1051] A "chat-based generative AI model" is a generative AI model for conducting text-based conversations.

[1052] An "image generation system generative AI model" is a generative AI model for generating images based on prompt sentences.

[1053] An "interface" is a means or method by which a user interacts with a system, and corresponds to the generative AI models of chat and image generation systems.

[1054] MODE FOR CARRYING OUT THE INVENTION

[1055] This invention relates to a system that receives input from a user, analyzes responses from multiple generative AI models, and generates and sends new prompt sentences. A specific embodiment of this system is described below.

[1056] Hardware and software used

[1057] The server uses the following hardware and software:

[1058] Hardware: A server machine with a powerful processor, sufficient memory, and a fast network connection.

[1059] Software: Operating systems (e.g., Linux), programming languages ​​(e.g., Python), natural language processing libraries (e.g., spaCy, NLTK), text matching algorithms (e.g., difflib, fuzzywuzzy), generative AI models (e.g., OpenAI's GPT-3, DALL-E).

[1060] System processing flow

[1061] A user uses a terminal to input a prompt to the system, such as "Generate an image of a cat" or "What is the diameter of the Earth?"

[1062] The server receives input from the user and sends prompts to multiple generative AI models using the APIs of the generative AI models (e.g., OpenAI's ChatGPT or DALL-E).

[1063] The server receives answers from each generative AI model, which may be in text or image format.

[1064] The server analyzes the received responses and extracts keywords using Python's natural language processing libraries, such as spaCy and NLTK. For example, in response to a prompt such as "Generate an image of a cat," the server extracts keywords such as "cat" and "image."

[1065] The server generates a new prompt based on the extracted keywords. This prompt is generated using a generative AI model (e.g., OpenAI's GPT-3). For example, it generates a new prompt such as "Generate a more detailed image of a cat."

[1066] The server compares answers from multiple generative AI models and, if there are identical answers, rates them as more reliable. This comparison uses text matching algorithms like "difflib" and "fuzzywuzzy."

[1067] The server sends the most credible answers to the terminal for display to the user, and the terminal shows the displayed text answers and image answers to the user.

[1068] Specific examples

[1069] Example 1: Receiving input from the user

[1070] A user types "generate an image of a cat" into a terminal.

[1071] The server receives this input and logs it.

[1072] Example 2: Sending prompts to multiple AIs

[1073] The server sends a prompt to ChatGPT and DALL-E APIs: "Generate a cat image."

[1074] The server checks the status of the API request and proceeds to the next step if successful.

[1075] Example 3: Receiving answers from AI

[1076] The server receives text responses from ChatGPT and image responses from DALL-E.

[1077] The server temporarily stores the received response.

[1078] Example 4: Answer analysis and keyword extraction

[1079] The server analyzes ChatGPT's text responses using "spaCy" and extracts keywords such as "cat" and "image."

[1080] The server records the extracted keywords in a log.

[1081] Example 5: Creating a new prompt

[1082] Based on the extracted keywords, the server generates a new prompt such as "Generate a more detailed image of a cat."

[1083] The server sends this prompt to the generative AI model.

[1084] Example 6: Judging credibility

[1085] The server uses "fuzzywuzzy" to compare answers received from multiple generative AI models, and if there are answers with the same content, it rates their credibility highly.

[1086] The server logs the results of the credibility assessment.

[1087] Example 7: Displaying results to the user

[1088] The server sends the most credible answer to the terminal for display to the user.

[1089] The terminal shows the displayed text and image answers to the user.

[1090] In this way, this system can effectively analyze responses from multiple generative AI models based on user input and generate new prompts. Furthermore, by determining the credibility of responses from multiple generative AI models, it is possible to provide highly reliable information to the user. Furthermore, by providing a unified interface for chat-based and image-based generative AI models, it is possible to ensure a consistent user experience. The flow of the identification process in Example 3 is described with reference to FIG. 15.

[1091] Step 1:

[1092] A user uses a terminal to input a prompt to the system, such as "Generate an image of a cat" or "What is the diameter of the Earth?"

[1093] Input: The prompt text entered by the user.

[1094] Output: The prompt received by the server.

[1095] Specific behavior: The user enters a prompt into the input field of the terminal and presses the send button. The server receives this input and records it in the log.

[1096] Step 2:

[1097] The server sends the received prompts to multiple generative AI models using the APIs of the generative AI models (e.g., OpenAI's ChatGPT or DALL-E).

[1098] Input: The prompt received by the server.

[1099] Output: The prompt sent to each generative AI model.

[1100] Specific operation: The server sends the prompt sentence as an API request to the generative AI model, checks the status of the API request, and proceeds to the next step if it is successful.

[1101] Step 3:

[1102] The server receives answers from each generative AI model, which may be in text or image format.

[1103] Input: The answer from each generative AI model.

[1104] Output: The answer stored on the server.

[1105] Specific operation: The server receives the API response from the generative AI model and temporarily stores the text and image answers.

[1106] Step 4:

[1107] The server analyzes the received responses and extracts keywords using Python's natural language processing libraries, spaCy and NLTK.

[1108] Input: The received text response.

[1109] Output: Extracted keywords.

[1110] Specific operation: The server analyzes the text response using "spaCy" and "NLTK" to extract keywords such as "cat" and "image." The extracted keywords are recorded in a log.

[1111] Step 5:

[1112] The server generates a new prompt based on the extracted keywords, using a generative AI model (e.g., OpenAI's GPT-3).

[1113] Input: Extracted keywords.

[1114] Output: The new prompt text.

[1115] Specific operation: The server generates a new prompt sentence based on the extracted keywords, such as "Generate a more detailed image of a cat," and sends this prompt sentence to the generative AI model.

[1116] Step 6:

[1117] The server compares answers from multiple generative AI models and, if there are identical answers, rates them as more reliable. This comparison uses text matching algorithms like "difflib" and "fuzzywuzzy."

[1118] Input: Answers from multiple generative AI models.

[1119] Output: The credibility assessment result.

[1120] Specific behavior: The server compares the received answers using "difflib" or "fuzzywuzzy", and if there are answers with the same content, it rates the answer as more reliable. It records the result of the reliability evaluation in a log.

[1121] Step 7:

[1122] The server sends the most credible answers to the terminal for display to the user, and the terminal shows the displayed text answers and image answers to the user.

[1123] Input: A credible answer.

[1124] Output: The answer that is displayed to the user.

[1125] Specific operation: The server sends a credible answer to the device, which displays it to the user. The user checks the displayed text and image answers.

[1126] (Application example 3)

[1127] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[1128] Conventional advertising material generation systems make it difficult for advertising creators to efficiently generate high-quality advertising copy and images. Furthermore, there was a lack of means to evaluate the credibility of responses from multiple AIs, which led to problems with the quality and reliability of the generated advertising materials. Furthermore, there was no interface compatible with both conversational AI and image-generating AI, making the advertising material generation process cumbersome.

[1129] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means. In this invention, the server includes means for sending the same instruction content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending new instructions based on the multiple received answers, means for determining the credibility of the same answer content, means for supporting dialogue-based AI and image-generating AI, means for generating instructions for generating advertising material, means for acquiring advertising copy and images from multiple generative AI models, means for evaluating the credibility of the acquired advertising copy and images, and means for displaying the acquired advertising copy and images. This enables advertising creators to efficiently generate high-quality, reliable advertising material.

[1130] "Multiple artificial intelligences" refers to multiple AI systems that operate using different algorithms and platforms.

[1131] "Instructions" refer to input information or commands that instruct the artificial intelligence to perform a specific task.

[1132] An "answer" is the output result generated by the artificial intelligence based on the instructions.

[1133] "Credibility" is the standard by which we evaluate whether the answers we receive are accurate and trustworthy.

[1134] "Conversational artificial intelligence" refers to AI systems that have the ability to converse with humans through text or voice.

[1135] "Image-generating artificial intelligence" refers to an AI system that has the ability to generate images based on input instructions.

[1136] "Advertising Materials" means content such as text and images used in an advertising campaign.

[1137] A "generative artificial intelligence model" is an AI model that generates content such as text and images for a specific task.

[1138] "Measures for assessing credibility" refers to methods or algorithms for assessing the accuracy and reliability of answers received from multiple AIs.

[1139] "Displaying means" refers to a method or device for visually presenting the generated advertising copy and images to the user.

[1140] The system for implementing this invention is a tool that allows advertising creators to efficiently generate high-quality advertising materials. The system has the function of sending the same instructions to multiple AIs, receiving responses from the multiple AIs, and creating and sending new instructions based on those responses. It also judges the credibility of the same responses and has an interface that supports dialogue-based AI and image-generating AI.

[1141] Hardware and software used

[1142] Hardware:

[1143] Smartphone

[1144] software:

[1145] Python

[1146] OpenAI API

[1147] PIL (Python Imaging Library)

[1148] requests

[1149] Data processing and calculation

[1150] Server Action:

[1151] 1. Prompt generation: Based on the keywords and themes entered by the ad creator, a prompt is generated to be sent to the generative AI model. For example, if the keyword "new smartphone release" is entered, the prompt will be as follows:

[1152] Ad theme: New smartphone launch. Generate the best copy and images.

[1153] 2. Obtaining answers from AI: Using OpenAI's API, we obtain ad copy and images from the generative AI model. Text answers are obtained from the conversational AI, and image answers are obtained from the image-generating AI.

[1154] 3. Credibility assessment: Compare the responses received to see if any of them are identical, thereby assessing their credibility.

[1155] 4. Displaying the results: Displaying the generated ad copy and fetching and displaying images. For example, if the generated ad copy is "This new smartphone is equipped with the latest technology" and the generated image is a high-resolution image of a smartphone, they will be visually presented to the ad creator.

[1156] Specific examples

[1157] When an ad creator enters the keyword "new smartphone release," the system generates the following prompt and sends it to the AI:

[1158] Ad theme: New smartphone launch. Generate the best copy and images.

[1159] The AI ​​generates the ad copy, "This new smartphone is equipped with the latest technology," along with a high-resolution image of the smartphone. The system evaluates these responses and, if deemed credible, displays them to the ad creator.

[1160] In this way, ad creators can efficiently generate high-quality, reliable ad materials.

[1161] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1162] Step 1:

[1163] The user inputs keywords or themes for generating advertising materials. For example, the input keyword might be "new smartphone release." This input becomes the basis for generating prompts.

[1164] Step 2:

[1165] The server generates a prompt based on the input keywords. Specifically, it creates the prompt "Advertisement theme: New smartphone release. Please generate the best ad copy and image." This prompt is used in the next step.

[1166] Step 3:

[1167] The server sends the generated prompt sentence to multiple generative AI models. In this case, the prompt sentence is sent to both conversational AI and image generation AI. The input is the prompt sentence, and the output is the answer from each AI model.

[1168] Step 4:

[1169] The server receives responses from multiple generative AI models. The conversational AI model returns ad copy, while the image-generating AI model returns ad images. For example, a high-resolution image of a smartphone might be generated along with the ad copy, "This new smartphone is equipped with the latest technology."

[1170] Step 5:

[1171] The server evaluates the credibility of the received answers. Specifically, it compares answers from multiple AIs to see if there are any answers with the same content. Answers that are judged to be highly credible are used in the next step.

[1172] Step 6:

[1173] The server displays the advertisement copy and images that are judged to be highly credible to the user. The user can review the generated advertisement copy and images and make corrections or adopt them as necessary. The output is the advertisement material presented to the user.

[1174] In this way, users can efficiently generate high-quality and reliable advertising materials.

[1175] 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.

[1176] "Example 1"

[1177] One embodiment of the present invention is a system incorporating an emotion engine. This system analyzes emotions from text and voice input from a user. Specifically, if a user inputs the text "Today was fun" to the system, the emotion engine recognizes the emotion "joy" from this text. Also, if a user inputs the text "Today was hard," the emotion engine recognizes the emotion "sadness."

[1178] "Example 2"

[1179] Furthermore, the emotion engine adjusts the AI's response based on the analyzed emotions. For example, if a user enters the text "I had fun today," and the emotion "joy" is recognized, the AI ​​will respond positively, such as "That's great! What was fun about it?" On the other hand, if a user enters the text "I had a hard day today," and the emotion "sadness" is recognized, the AI ​​will respond with an empathetic response, such as "That must have been tough. Maybe talking about it will help you feel a little better."

[1180] "Example 3"

[1181] In this way, a system incorporating an emotion engine can properly recognize the user's emotions and generate AI responses accordingly, enabling more human-like communication, which makes the user's interaction with the system feel more natural and satisfying.

[1182] The processing flow of each embodiment will be described below.

[1183] "Example 1"

[1184] Step 1: The user inputs text into the system. For example, the user inputs the text "I had a great time today."

[1185] Step 2: The emotion engine analyzes the emotion from the user's text input. In this example, the emotion "joy" is recognized.

[1186] "Example 2"

[1187] Step 1: The user inputs text into the system, for example, "Today was a tough day."

[1188] Step 2: The emotion engine analyzes the emotion from the user's text input. In this example, the emotion "sadness" is recognized.

[1189] Step 3: The emotion engine adjusts the AI's response based on the analyzed emotions. In this example, it generates an empathetic response such as, "That must be tough. Maybe talking about it will help you feel a little better."

[1190] "Example 3"

[1191] Step 1: The user provides text input to the system.

[1192] Step 2: The emotion engine analyzes the emotion from the user's text input.

[1193] Step 3: The emotion engine adjusts the AI's response based on the analyzed emotions.

[1194] Step 4: The AI-generated response is returned to the user, making the user's interaction with the system feel more natural and satisfying.

[1195] Example 1

[1196] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1197] In conventional systems, it was difficult to send the same prompt to multiple AIs and collect their responses. Furthermore, the system lacked the ability to analyze the user's emotions and generate responses based on them, making it impossible to provide appropriate responses that met the user's needs. Furthermore, there was also an insufficient means to determine the credibility of responses from multiple AIs.

[1198] 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.

[1199] In this invention, the server includes means for sending the same prompt content to multiple AIs, means for receiving answers from the multiple AIs, means for creating and sending new prompts based on the received multiple answers, means for determining the credibility of the same answer content, means for analyzing emotions from user input, means for generating responses based on the analyzed emotions, and means for supporting chat-type and image-generating AIs. This makes it possible to efficiently collect responses from the multiple AIs and provide appropriate responses based on the user's emotions.

[1200] "Artificial intelligence" is the technology that enables computer systems to learn, reason, and self-correct by imitating human intelligence.

[1201] A "prompt" is an input sentence that causes an artificial intelligence to generate a particular response.

[1202] "Emotion analysis" is a technology that identifies emotions from user input and evaluates the type and intensity of those emotions.

[1203] "Credibility" is the standard by which information or responses are judged to be accurate and trustworthy.

[1204] "Chat AI" is AI designed for text-based conversations.

[1205] "Image generation AI" is AI designed to generate images based on input data.

[1206] A "platform" is the underlying hardware and software environment for running software and services.

[1207] A "network" is a system that allows multiple computers and devices to communicate with each other.

[1208] "Merchandise" refers to goods and services that are the subject of commercial activity.

[1209] This invention is a system that sends the same prompt content to multiple artificial intelligences, collects their responses, analyzes the user's emotions, and generates an appropriate response. Specific embodiments of this system will be described below.

[1210] The server is connected to a terminal for receiving prompt sentences from the user. When the user inputs a prompt sentence using the terminal, the prompt sentence is sent to the server. The server passes the received prompt sentence to an emotion engine, which analyzes the user's emotion. The emotion engine identifies emotions from the user's input text and evaluates the type and intensity of the emotion.

[1211] The server then sends prompts to multiple AIs based on the analyzed emotional information. Specifically, the server uses the API of the cloud platform to send the same prompt to different AIs. For example, the server uses the API of cloud service A to send a prompt to AI model A, and simultaneously sends the same prompt to AI model B using the API of cloud service B.

[1212] Each AI model generates a response to the received prompt. The generated response is sent to the server. The server receives responses from multiple AI models and judges the credibility of those responses. The credibility judgment takes into account the degree of similarity of the response content and past response history.

[1213] Finally, the server returns a highly credible response to the user. For example, if a user inputs a prompt sentence such as "What's the weather like today?", the server receives a response such as "It's sunny today" from AI model A of cloud service A, and a response such as "It's cloudy today" from AI model B of cloud service B. The server compares these responses and returns a highly credible response to the user.

[1214] As a concrete example, the processing when the user inputs the prompt sentence "What's the weather like today?" is shown below.

[1215] 1. A user uses a terminal to type, "What's the weather like today?"

[1216] 2. The server receives the prompt and performs emotion analysis using the emotion engine.

[1217] 3. The server uses the API of cloud service A to send a prompt to artificial intelligence model A, and simultaneously sends the same prompt to artificial intelligence model B using the API of cloud service B.

[1218] 4. AI model A responds, "It's sunny today," and AI model B responds, "It's cloudy today."

[1219] 5. The server receives these responses and determines their authenticity.

[1220] 6. The server returns a reliable response to the user.

[1221] In this way, the server can efficiently collect responses from multiple artificial intelligences and provide an appropriate response based on the user's emotions.

[1222] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1223] Step 1:

[1224] The user enters a prompt statement.

[1225] The user uses a terminal to input a prompt to the system, for example, "What's the weather like today?" The input prompt is sent from the terminal to the server.

[1226] Input: User prompt text

[1227] Output: The prompt sent to the server

[1228] Step 2:

[1229] The server receives the prompt.

[1230] The server receives the prompt sent by the user, which becomes the input data for the next processing step.

[1231] Input: Prompt sent from the terminal

[1232] Output: Received prompt text

[1233] Step 3:

[1234] The server performs emotion analysis using an emotion engine.

[1235] The server passes the received prompt sentence to the emotion engine for emotion analysis. For example, if the user enters "I had fun today," the emotion engine recognizes the emotion of "joy."

[1236] Input: Received prompt text

[1237] Output: Parsed emotion information

[1238] Step 4:

[1239] The server sends prompt sentences to multiple artificial intelligences.

[1240] The server uses the API of the cloud platform to send the same prompt to different AIs, for example, to send a prompt to AI model A using the API of cloud service A, and at the same time to send the same prompt to AI model B using the API of cloud service B.

[1241] Input: Received prompt sentence, parsed emotion information

[1242] Output: Prompt sent to each AI

[1243] Step 5:

[1244] Each AI generates a response to the prompt.

[1245] AI model A and AI model B each generate a response to the received prompt. For example, AI model A responds, "It's sunny today," and AI model B responds, "It's cloudy today."

[1246] Input: Prompt sent to each AI

[1247] Output: The generated response

[1248] Step 6:

[1249] The server receives responses from each artificial intelligence.

[1250] The server receives responses from AI model A and AI model B. This allows information from multiple AIs to be aggregated on the server.

[1251] Input: Responses from each AI

[1252] Output: Response received

[1253] Step 7:

[1254] The server determines the authenticity of the response.

[1255] The server determines the authenticity of the received response, taking into account the degree of match of the response content and past response history.

[1256] Input: Received response

[1257] Output: A credible response

[1258] Step 8:

[1259] The server returns the response to the user.

[1260] The server responds to the user with a highly credible response, for example, "OpenAI's response: It's sunny today. Bard's response: It's cloudy today."

[1261] Input: A credible response

[1262] Output: The response returned to the user

[1263] (Application example 1)

[1264] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1265] Conventional AI systems have had difficulty analyzing user emotions and generating optimal prompts based on those emotions. It has also been difficult to send the same prompt to multiple AIs and provide optimal customer service methods and product suggestions in real time based on their responses. This has made it difficult to improve customer satisfaction in brick-and-mortar stores.

[1266] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means. In this invention, the server includes means for sending the same prompt content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending a new prompt based on the multiple received answers, means for determining the credibility of the same answer content, means for supporting chat-type and image-generation-type AIs, means for analyzing a user's emotions using an emotion analysis engine, means for generating an optimal prompt based on the analyzed emotions, and means for sending the generated prompt to multiple AIs and providing the obtained responses. This makes it possible to provide optimal customer service methods and product suggestions in real time according to the user's emotions.

[1267] "Multiple artificial intelligences" refers to multiple AI systems that operate using different algorithms and platforms.

[1268] A "prompt" is text containing instructions or questions that instruct an artificial intelligence to perform a specific task.

[1269] An "emotion analysis engine" is software that analyzes emotions from a user's text or voice input and identifies those emotions.

[1270] A "means for determining credibility" is an algorithm or mechanism for assessing the accuracy and reliability of answers obtained from multiple artificial intelligences.

[1271] "Chat AI" refers to an AI system designed for text-based dialogue.

[1272] "Image generation artificial intelligence" is an AI system that generates images based on input data.

[1273] A "means for generating optimal prompts" is an algorithm or mechanism for generating optimal instructions or questions for the artificial intelligence based on analyzed emotions and other data.

[1274] A "means for providing a response" is a mechanism for presenting answers obtained from multiple artificial intelligences to a user or system.

[1275] The system for implementing this invention sends the same prompt content to multiple artificial intelligences, and based on the responses obtained, provides optimal customer service methods and product suggestions in accordance with the user's emotions. A specific embodiment of this system will be described below.

[1276] System Configuration

[1277] The system consists of the following main components:

[1278] 1. Sentiment Analysis Engine: Software that analyzes emotions from a user's text or voice input.

[1279] 2. Prompt generation engine: Software that generates optimal prompts based on the analyzed emotions.

[1280] 3. Multiple Artificial Intelligence: Different AI systems connected via cloud services (e.g., generative AI models via cloud services).

[1281] 4. Credibility Judgment Engine: Software that evaluates the credibility of answers obtained from multiple artificial intelligences.

[1282] 5. Response Delivery Engine: Software that delivers the resulting responses to users or systems.

[1283] Hardware and Software

[1284] The system uses the following hardware and software:

[1285] Smart glasses: A device worn by store clerks to analyze conversations and behavior with customers in real time.

[1286] Cloud Services: Platforms (e.g., cloud services) for hosting sentiment analysis engines and generative AI models.

[1287] Python: A programming language for implementing the system's programs.

[1288] Data processing and calculation

[1289] 1. Sentiment analysis: The user's text or voice input is sent to the sentiment analysis engine to analyze their emotions. For example, if the user says, "Today was fun," the sentiment analysis engine will recognize the emotion of "joy."

[1290] 2. Prompt generation: Based on the analyzed sentiment, the prompt generation engine generates the most appropriate prompt. For example, it generates a prompt such as, "What is the best way to serve a user when they are happy?"

[1291] 3. Prompt sending: Send the generated prompt to multiple AIs and get responses from each AI.

[1292] 4. Credibility judgment: The credibility of the obtained response is evaluated by the credibility judgment engine.

[1293] 5. Response provision: The response provision engine provides the evaluated responses to the store staff, who then suggest optimal customer service methods and products in real time.

[1294] Specific examples

[1295] For example, if a customer says, "I had a great time today," the emotion analysis engine recognizes the emotion of "joy." The prompt generation engine generates a prompt, "Please tell me the best way to serve a customer when they are happy," and sends it to multiple AIs. The credibility judgment engine evaluates the responses obtained and provides the optimal response to the store clerk.

[1296] Prompt Sentence Examples

[1297] "What is the best way to serve a user when they are happy?"

[1298] In this way, the system can provide optimal customer service methods and product suggestions in real time based on the user's emotions.

[1299] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1300] Step 1:

[1301] The user inputs text or voice. For example, the user says, "I had fun today." The input data is sent to the server through the smart glasses.

[1302] Step 2:

[1303] The server sends the received user input data to the emotion analysis engine. The emotion analysis engine analyzes the input data and identifies the user's emotion. For example, it recognizes the emotion "joy" from the text "I had fun today." The input is the user's text data, and the output is analyzed emotion data.

[1304] Step 3:

[1305] The server uses a prompt generation engine based on the analyzed emotion data to generate an optimal prompt. For example, for the emotion of "happiness," it generates the prompt "Please tell me the best way to serve a user when they are happy." The input is emotion data, and the output is the generated prompt.

[1306] Step 4:

[1307] The server sends the generated prompts to multiple artificial intelligences, for example, multiple generative AI models connected via a cloud service. The input is the generated prompt, and the output is the response data from each artificial intelligence.

[1308] Step 5:

[1309] The server sends the response data obtained from multiple AIs to the credibility judgment engine. The credibility judgment engine evaluates the accuracy and reliability of each response. The input is the response data, and the output is the evaluated credibility data.

[1310] Step 6:

[1311] The server selects the optimal response based on the credibility data obtained from the credibility judgment engine. The selected response is sent to the smart glasses through the response providing engine and provided to the store clerk. The input is the credibility data, and the output is the optimal response data.

[1312] Step 7:

[1313] The terminal (smart glasses) displays the optimal response data received from the server to the store clerk. Based on the displayed response, the store clerk will provide the user with the optimal customer service method and product suggestions. The input is the optimal response data, and the output is the store clerk's behavior.

[1314] In this way, the system can provide optimal customer service methods and product suggestions in real time according to the user's emotions.

[1315] Example 2

[1316] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1317] Conventional AI systems have difficulty managing responses from multiple AIs in a unified manner and providing appropriate responses based on the user's emotions. Furthermore, they lacked the means to integrate AIs provided via different platforms and select highly credible answers. Furthermore, they lacked the ability to analyze the user's emotions and adjust responses based on them, limiting the user experience.

[1318] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1319] In this invention, the server includes means for sending the same instruction content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending new instructions based on the multiple received answers, means for determining the credibility of the same answer content, means for supporting conversational and image generation AIs, means for receiving user input and analyzing emotions with an emotion analysis engine, means for adjusting the AI's response based on the analyzed emotions, means for saving the received answers in a database, means for selecting an optimal response based on emotions, and means for returning a response to the user. This makes it possible to centrally manage answers from multiple AIs and provide appropriate responses according to the user's emotions.

[1320] "Artificial intelligence" is the technology that enables computer systems to learn, reason, and self-correct by imitating human intelligence.

[1321] "Instructions" are the information or instructions given to an artificial intelligence to perform a specific task or question.

[1322] An "answer" is a response or result generated by artificial intelligence based on instructions.

[1323] "Credibility" is a criterion for assessing whether a response is accurate and trustworthy.

[1324] "Conversational AI" is AI that uses natural language processing to engage in dialogue with users.

[1325] "Image generation artificial intelligence" is artificial intelligence that generates images based on input data.

[1326] An "emotion analysis engine" is software that analyzes emotions from text entered by a user and recognizes those emotions.

[1327] A "database" is a system for efficiently storing, managing, and retrieving data.

[1328] An "optimal response" is an answer that is provided in a manner that best suits the user's feelings and situation.

[1329] A "platform" is the underlying environment or infrastructure for providing software and services.

[1330] This invention is a system that centrally manages responses from multiple artificial intelligences and provides appropriate responses according to the user's emotions. Specific embodiments of this system will be described below.

[1331] The server has a means to send the same instructions to multiple AIs. This is achieved by sending API requests using the Python requests library. For example, if a user enters "I had fun today," the server sends this text as a prompt to multiple AIs.

[1332] Next, the server has a means to receive answers from multiple AIs. This is also realized by receiving API responses using Python's requests library. For example, it receives a response from AI1 saying, "That's great! What was fun about it?"

[1333] The server also includes a means for generating and sending new instructions based on the received responses. This means includes logic for analyzing the received responses and generating the next prompt. For example, if the user inputs "Today was a bad day," the server generates a new prompt that reads, "The user is feeling sad. Please respond empathetically."

[1334] The server also includes a means for determining the authenticity of the content of the identical answers, the means including an algorithm for comparing the content of the received answers and assessing the authenticity of matching answers.

[1335] Furthermore, the server has the means to support conversational and image generation artificial intelligence, which enables not only text-based conversation but also image generation.

[1336] The server also includes a means for receiving user input and analyzing emotions using a sentiment analysis engine. This means uses natural language processing libraries such as NLTK and Transformers to analyze text and recognize emotions. For example, it recognizes the emotion "joy" from the text "I had fun today."

[1337] The server also includes a means for adjusting the response of the artificial intelligence based on the analyzed emotions, thereby enabling it to generate an appropriate response according to the user's emotions.

[1338] The server also contains a mechanism for storing the received answers in a database, which uses the Python pymysql library to connect to a MySQL database and store the answers.

[1339] The server also includes a means for selecting an optimal response based on emotion, the means including logic for selecting an optimal response from among the stored answers based on the analysis of the emotion engine.

[1340] Finally, the server has a means for returning a response to the user. This means includes logic for sending the selected response to the user's device. For example, if the user enters "I had fun today," the server will send the response "That's great! What did you enjoy?" to the user's device.

[1341] For example, if the user types "Today was a tough day," here's the prompt:

[1342] User: "Today was fun"

[1343] AI: "That's great! What was fun about it?"

[1344] User: "Today was tough."

[1345] AI: "That must have been tough. Maybe talking about it will make you feel a little better."

[1346] In this way, the server centrally manages responses from multiple AIs, creating a system that provides appropriate responses according to the user's emotions.

[1347] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1348] Step 1:

[1349] Accepting user input

[1350] A user inputs text into the system using a terminal. For example, the user inputs "I had a good time today." This input text is sent to the server.

[1351] Step 2:

[1352] Analyzing emotions with an emotion engine

[1353] The server receives the user's input text and passes it to the sentiment analysis engine. The sentiment analysis engine uses a natural language processing library (such as NLTK or Transformers) to analyze the text and recognize emotions. For example, it can recognize the emotion "joy" from the text "I had fun today." The analysis results are then returned to the server.

[1354] Step 3:

[1355] Send prompts to each AI

[1356] The server sends prompts to the generative AI model based on the analyzed emotions. For example, it sends a prompt such as "The user feels happy. Please ask him what he enjoyed." This prompt is then sent to each AI.

[1357] Step 4:

[1358] Receive answers from each AI

[1359] The server receives the answers from each AI through the API. For example, AI1 will respond, "That's great! What was fun about it?" These answers are then aggregated on the server.

[1360] Step 5:

[1361] Save the answers to the database

[1362] The server stores the received answers in a MySQL database. It uses the Python pymysql library to connect to the database and store the answers. For example, the answer "That's great! What did you enjoy?" is stored in the database.

[1363] Step 6:

[1364] Choose the best response based on emotion

[1365] The server selects the best response from the stored answers based on the analysis results of the emotion engine. For example, it selects the most appropriate response for the emotion of "joy." This selected response is stored on the server.

[1366] Step 7:

[1367] Returning a response to the user

[1368] The server sends the selected response to the user's device, where the user receives the response "That's great! What did you enjoy?" This response is displayed on the user's device.

[1369] (Application example 2)

[1370] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1371] Conventional AI systems have had difficulty providing appropriate responses and recommending content based on user emotions. In particular, there was no system that could centrally manage responses from multiple AIs and adjust responses based on user emotions. This could lead to inconsistent user experiences and reduced satisfaction.

[1372] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1373] In this invention, the server includes means for sending the same prompt content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending new prompts based on the received multiple answers, means for determining the credibility of the same answer content, means for supporting chat-type and image-generating AIs, means for analyzing emotions from user input text, means for adjusting the AI's response based on the analyzed emotions, and means for recommending content according to the emotions. This makes it possible to recommend appropriate responses and content according to the user's emotions.

[1374] "Multiple AIs" refers to multiple artificial intelligence systems that operate using different algorithms and data sets.

[1375] A "prompt" is an instruction or question given to an AI to help it perform a specific task.

[1376] An "answer" is the response or output that the AI ​​generates based on a prompt.

[1377] "Credibility" is a criterion for assessing whether answers from AI are accurate and trustworthy.

[1378] "Chat AI" refers to an artificial intelligence system designed for text-based dialogue.

[1379] "Image generation AI" refers to an artificial intelligence system designed to generate images.

[1380] "Means for emotion analysis" refers to technology that identifies emotions from user input text and evaluates the type and intensity of those emotions.

[1381] "Means to adjust responses" refers to technology that changes the AI's responses based on analyzed emotions.

[1382] "Content recommendation means" refers to technology that suggests appropriate content such as movies, music, and articles based on the user's emotions and preferences.

[1383] An embodiment of this invention will now be described. First, the entire system is composed of a server, a user terminal, and multiple AIs. The server sends the same prompt content to multiple AIs and receives answers from them. It creates and sends a new prompt based on the multiple answers it receives. It also has the function of judging the credibility of the same answer content. Furthermore, it is compatible with chat-type and image-generation-type AIs.

[1384] The user terminal is a device such as a smartphone or PC, which provides an interface for users to input text. The text entered by the user is sent to the server, which analyzes the text for emotion. For emotion analysis, an emotion analysis library such as TextBlob is used. Based on the analyzed emotion, the server adjusts the AI's response and recommends appropriate content.

[1385] The specific hardware and software configuration is as follows: the server uses Flask (a Python web framework) and the TextBlob library for sentiment analysis. User devices are general devices such as smartphones and PCs, and communicate with the server via an internet connection.

[1386] For example, if a user types "I had fun today," the server receives the text and analyzes it for the emotion of "joy" using the TextBlob library. The server then recommends fun movies and upbeat music based on the emotion of "joy." An example of a specific prompt sentence is the text "I had fun today."

[1387] In this way, it becomes possible to provide appropriate responses and recommend content according to the user's emotions.

[1388] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1389] Step 1:

[1390] The user enters text into the terminal.

[1391] Input: Text entered by the user (e.g., "I had fun today")

[1392] Output: Text data sent from the terminal to the server

[1393] Step 2:

[1394] The server passes the received text to a sentiment analysis engine.

[1395] Input: Text data sent from the terminal

[1396] Output: Text data passed to the sentiment analysis engine

[1397] Step 3:

[1398] The server analyzes the sentiment of the text using a sentiment analysis engine.

[1399] Input: Text data passed to the sentiment analysis engine

[1400] Data Processing: Sentiment Analysis using the TextBlob Library

[1401] Output: Parsed emotion data (e.g., "joy")

[1402] Step 4:

[1403] The server sends prompts to multiple AIs based on the analyzed emotions.

[1404] Input: Parsed emotion data

[1405] Data processing: Generate prompts based on sentiment data

[1406] Output: Prompt data sent to multiple AIs

[1407] Step 5:

[1408] Multiple AIs generate answers based on the prompts and send them to the server.

[1409] Input: Prompt data sent from the server

[1410] Data calculations: Each AI generates an answer based on a prompt

[1411] Output: Answer data sent from the AI ​​to the server

[1412] Step 6:

[1413] The server receives answers from multiple AIs and judges their credibility.

[1414] Input: Answer data sent from AI

[1415] Data calculation: Evaluating the degree of agreement and reliability of answers

[1416] Output: Answer data with credibility assessed

[1417] Step 7:

[1418] The server recommends content that matches the user's emotions based on the most reliable answers.

[1419] Input: Answer data with credibility assessment

[1420] Data processing: Select content based on sentiment data and response data

[1421] Output: Recommended content data (e.g., "fun movies," "upbeat music")

[1422] Step 8:

[1423] The server transmits the recommended content to the user terminal.

[1424] Input: Recommended content data

[1425] Output: Content data sent to the user device

[1426] Step 9:

[1427] The user terminal displays the received content to the user.

[1428] Input: Content data sent from the server

[1429] Output: The content displayed to the user (e.g., "Fun movie," "Upbeat music")

[1430] Example 3

[1431] Next, a description will be given of Example 3 of Form Example 3. 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."

[1432] In conventional AI systems, responses to user input often lacked credibility and specificity, and there were limited ways to effectively utilize responses from multiple AIs. Furthermore, there was a lack of interfaces that supported both chat AI and image generation AI, limiting the quality and variety of information available to users.

[1433] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1434] In this invention, the server includes means for receiving user input, means for sending the received input as a prompt to the generative AI model, means for receiving answers from the generative AI model, means for analyzing answers from the generative AI models, means for extracting keywords from the analyzed answers and generating a new prompt, means for resending the new prompt to the generative AI model, means for evaluating the credibility of answers from the generative AI models, and means for displaying answers from the chat AI and the image generation AI. This allows users to obtain highly reliable information and further enables them to effectively utilize both text information and image information.

[1435] The "means for receiving user input" is an interface through which the user inputs questions or requests to the system.

[1436] "Means for sending received input as a prompt to the generative AI model" is a function for sending input from the user to the generative AI model in an appropriate format.

[1437] "Means for receiving an answer from a generative AI model" is a function for receiving an answer returned from a generative AI model.

[1438] "Means for analyzing answers from multiple generative AI models" refers to a function for comparing and analyzing answers obtained from multiple generative AI models.

[1439] "Means of extracting keywords from analyzed answers and generating new prompts" is a function that extracts important keywords from the analysis results and creates new prompts based on them.

[1440] The "means for resending a new prompt to the generation AI model" is a function for resending the generated new prompt to the generation AI model.

[1441] "Means for evaluating the credibility of answers from multiple generative AI models" is a function for evaluating the reliability of answers obtained from multiple generative AI models.

[1442] "Means for displaying answers from chat AI and image generation AI" is a function for appropriately displaying text answers from chat AI and image answers from image generation AI to the user.

[1443] MODE FOR CARRYING OUT THE INVENTION

[1444] This invention is a system that receives user input, sends prompts to a generative AI model based on that input, analyzes the responses from the generative AI model, and generates new prompts. It also has the function of evaluating the credibility of responses from multiple generative AI models and appropriately displaying the responses of the chat AI and image generation AI.

[1445] Hardware and software used

[1446] The system uses the following hardware and software:

[1447] Server: A central processing unit for receiving user input, sending prompts to the generative AI model, and analyzing the answers.

[1448] Terminal: A device used by a user to input information (e.g., a computer or smartphone)

[1449] Generative AI model: An artificial intelligence model (e.g., GPT-4) for generating answers to prompts.

[1450] Chat AI: Artificial intelligence that generates text-based responses (e.g., ChatGPT)

[1451] Image generation AI: Artificial intelligence that generates answers in image format (e.g., DALL-E)

[1452] Data processing and calculation

[1453] The server receives input from the user and sends it as a prompt to the generative AI model. After receiving the answer from the generative AI model, the server analyzes the answers from multiple generative AI models and extracts keywords. It generates a new prompt based on the extracted keywords and sends it again to the generative AI model. It then compares the answers from multiple generative AI models and evaluates their credibility. Finally, the answers from the chat AI and image generation AI are displayed appropriately on the user's device.

[1454] Specific examples

[1455] For example, if a user types "Please tell me your ideas for a new project" on their device, the server sends this input as a prompt to the generative AI model. The generative AI model responds with a response such as "One option is developing environmentally friendly products." The server analyzes this response and extracts the keywords "environmentally friendly products" and "development." The server then generates a new prompt, "Please tell me specific methods for developing environmentally friendly products," and sends it again to the generative AI model. If the generative AI model responds with a response such as "Using recycled materials is effective," the server compares the responses from multiple generative AI models and evaluates their credibility. Finally, the server displays the text response from the chat AI and the image response from the image generation AI on the user's device.

[1456] Prompt Sentence Examples

[1457] "Tell me your ideas for a new project."

[1458] "Please tell me the specific methods for developing environmentally friendly products."

[1459] "What factors contributed to the success of this project?"

[1460] "What are the risks of the project?"

[1461] In this way, the user can obtain specific and highly reliable information through the system. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1462] Step 1:

[1463] The user types at the terminal.

[1464] Specific Action: The user uses the device interface to enter a question or request, for example, "Can you give me some ideas for a new project?"

[1465] Input: User question or request

[1466] Output: User input data

[1467] Step 2:

[1468] The server sends the user's input as a prompt to the generative AI model.

[1469] Specific operation: The server receives the user's input and sends it to the generative AI model as a prompt sentence.

[1470] Input: User-entered data

[1471] Output: The prompt sent to the generative AI model

[1472] Step 3:

[1473] The generative AI model generates answers to the prompts and returns them to the server.

[1474] How it works: The generative AI model generates an answer based on the prompt it receives and returns the answer to the server, such as "We can consider developing environmentally friendly products."

[1475] Input: A prompt sent to the generative AI model

[1476] Output: Answer from the generative AI model

[1477] Step 4:

[1478] The server analyzes the answers from multiple generative AI models.

[1479] How it works: The server receives answers from multiple generative AI models and compares and analyzes them. During analysis, the content of the answers is compared to find similarities and differences.

[1480] Input: Answers from multiple generative AI models

[1481] Output: Analysis results (commonalities and differences)

[1482] Step 5:

[1483] The server extracts keywords from the analyzed answers and generates a new prompt.

[1484] Specific operation: The server extracts important keywords from the analysis results and creates a new prompt based on them. For example, the keywords extracted are "environmentally friendly products" and "development."

[1485] Input: Analysis results

[1486] Output: New prompt text

[1487] Step 6:

[1488] The server resends a new prompt to the generation AI model.

[1489] Specific operation: The server sends the newly generated prompt to the generation AI model again, for example, "Please tell me specific ways to develop environmentally friendly products."

[1490] Input: New prompt text

[1491] Output: The prompt sentence resubmitted to the generative AI model

[1492] Step 7:

[1493] The generative AI model generates answers to new prompts and returns them to the server.

[1494] How it works: The generative AI model generates an answer based on the new prompt and returns it to the server, for example, "Using recycled materials is effective."

[1495] Input: The prompt sentence resubmitted to the generative AI model

[1496] Output: A new answer from the generative AI model

[1497] Step 8:

[1498] The server evaluates the credibility of answers from multiple generative AI models.

[1499] Specific operation: The server compares answers from multiple generative AI models, and if there are answers with the same content, it rates their credibility higher.

[1500] Input: Answers from multiple generative AI models

[1501] Output: Credibility evaluation result

[1502] Step 9:

[1503] The server displays the answers from the chat AI and image generation AI.

[1504] Specific operation: The server receives both the text response from the chat AI and the image response from the image generation AI, and displays them appropriately on the user's device. For example, the text response might say "Use recycled materials," and the image response might show an image of a product made from recycled materials.

[1505] Input: Credibility assessment results, text response from chat AI, image response from image generation AI

[1506] Output: Text and image answers displayed on the user's device

[1507] (Application example 3)

[1508] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[1509] In modern brick-and-mortar stores, it is difficult for users to quickly and accurately obtain detailed information about products. There is also a need to provide a more natural and satisfying shopping experience by providing responses that reflect the user's emotions. Furthermore, there is a lack of a way to integrate information from multiple information processing devices and provide highly reliable information.

[1510] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1511] In this invention, the server includes means for sending the same instruction content to a plurality of information processing devices, means for receiving responses from the plurality of information processing devices, means for creating and sending new instructions based on the received plurality of responses, means for determining the credibility of the same response content, means for supporting both interactive information processing devices and image generation information processing devices, means for recognizing a user's emotions and generating a response accordingly, means for generating a new question in response to a user's question based on responses from the plurality of information processing devices and providing detailed information, means for analyzing the user's facial expression and tone of voice, means for analyzing information from the plurality of information processing devices and providing highly credible information when the user scans an item, and means for generating images of clothes the user wants to try on and providing a try-on simulation. This allows users to quickly and accurately obtain information in a physical store and receive responses according to their emotions, enabling a more natural and satisfying shopping experience.

[1512] An "information processing device" is a device for processing and analyzing data.

[1513] The "instruction content" is the content that specifies the processing or operation that should be performed by the information processing device.

[1514] A "response" is information that an information processing device returns based on the contents of an instruction.

[1515] "Credibility" is the degree to which information or responses are accurate and trustworthy.

[1516] An "interactive information processing device" is a device that processes information through interaction with a user.

[1517] An "image generation type information processing device" is an information processing device for generating images.

[1518] The "means for recognizing emotions" refers to a method or device for analyzing and recognizing a user's emotions.

[1519] A "means for generating new questions" is a method or device for creating new questions based on the responses received.

[1520] "Detailed information" is specific and in-depth information about a particular subject.

[1521] The "means for analyzing facial expressions and tone of voice" refers to a method or device for analyzing a user's facial expressions and tone of voice.

[1522] "Means for scanning products" refers to a method or device for reading product information.

[1523] "Try-on simulation" involves generating images of the clothes the user wants to try on, providing the experience of virtually trying them on.

[1524] A system for implementing this invention has the functions of sending the same instruction content to multiple information processing devices, receiving responses from the multiple information processing devices, and creating and sending new instructions based on the responses.It also has the functions of determining the authenticity of the same response content and supporting both interactive information processing devices and image generation information processing devices.It also includes the functions of recognizing the user's emotions and generating responses accordingly.

[1525] Hardware and software used

[1526] Hardware: Camera (to capture the user's facial expressions), smartphone (as a user interface)

[1527] software:

[1528] OpenAI API: To get answers from AI

[1529] EmotionRecognizer: To recognize user emotions

[1530] OpenCV: for processing camera images

[1531] System Operation

[1532] The server first captures the user's facial expression using a camera and recognizes their emotions using EmotionRecognizer. Next, when the user enters a question about a product via their smartphone, the server sends the question to multiple information processing devices using the OpenAI API. After receiving responses from the information processing devices, the server analyzes the responses and selects the most reliable information.

[1533] The server then extracts keywords from the response and generates a new prompt. Based on this new prompt, the server sends a question to the information processing device to obtain more detailed information. The obtained information is then displayed on the smartphone in a format that reflects the user's emotions.

[1534] Specific examples

[1535] For example, if a user asks, "I'd like to know more about this product," the server generates the following prompt and sends it to the information processing device:

[1536] Example prompt sentence:

[1537] A user has asked, "I want to know more about this product." Please provide the following information:

[1538] Product details

[1539] Other users' reviews

[1540] How to use

[1541] Related products

[1542] After receiving the responses from the information processing devices, the server analyzes the responses and selects information with high credibility. For example, if the server receives responses such as "This product is of high quality" from multiple information processing devices, it will rate the credibility of that information as high.

[1543] Furthermore, when a user generates an image of the clothes they want to try on, the server uses an image generation information processing device to provide a try-on simulation, allowing the user to virtually try on the clothes without actually trying them on.

[1544] In this way, the system provides users with fast and accurate information and generates emotionally appropriate responses, resulting in a more natural and satisfying shopping experience.

[1545] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1546] Step 1:

[1547] The server captures the user's facial expressions through a camera. The input is the camera image, and the output is the captured image data. Specifically, the server acquires the image from the camera in real time, divides the image into frames, and saves them as image data.

[1548] Step 2:

[1549] The server uses EmotionRecognizer to recognize the user's emotions from the captured image data. The input is image data, and the output is recognized emotion data. Specifically, the server inputs the image data into EmotionRecognizer and applies the emotion recognition algorithm to analyze the user's emotions.

[1550] Step 3:

[1551] A user inputs a question about a product through a smartphone. The input is the user's question text, and the output is the question data sent to the server. In concrete terms, the user inputs a question using a smartphone application, and the question is sent to the server.

[1552] Step 4:

[1553] The server uses the OpenAI API to send user questions to multiple information processing devices. The input is the question data, and the output is the instructions sent to the information processing devices. Specifically, the server inputs the question data into the OpenAI API and sends the same instructions to multiple information processing devices.

[1554] Step 5:

[1555] The server receives responses from multiple information processing devices. The input is the response data from the information processing devices, and the output is the received response data. Specifically, the server collects responses from each information processing device through the OpenAI API and saves them as a single dataset.

[1556] Step 6:

[1557] The server analyzes the received response data and selects the most credible information. The input is the response data, and the output is the most credible information. Specifically, the server inputs the response data into an analysis algorithm, and if there are multiple responses with the same content, it rates the most credible one.

[1558] Step 7:

[1559] The server extracts keywords from the received response and generates a new prompt. The input is credible information, and the output is a new prompt. Specifically, the server extracts important keywords from the credible information and creates a new prompt based on them.

[1560] Step 8:

[1561] The server then sends a new question to the information processing device based on the new prompt and obtains detailed information. The input is the new prompt text, and the output is the detailed information. Specifically, the server inputs the new prompt text into the OpenAI API and obtains detailed information from the information processing device.

[1562] Step 9:

[1563] The server then displays the acquired detailed information on the smartphone in a format that corresponds to the user's emotions. The input is detailed information and emotional data, and the output is a response that is displayed on the smartphone. Specifically, the server combines the detailed information and emotional data to generate a response that is optimal for the user and sends it to the smartphone.

[1564] Step 10:

[1565] The user checks the information displayed on their smartphone and asks additional questions or takes action as necessary. The input is the response displayed on the smartphone, and the output is the user's next action. Specifically, the user looks at the smartphone screen to check the information, and if necessary, enters additional questions or purchases a product.

[1566] 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.

[1567] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by 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.

[1568] Another example of generative AI is Gemini (internet search engine). <url: https: gemini.google.com ?hl="ja">) are mentioned.

[1569] 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.

[1570] [Third embodiment]

[1571] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1572] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[1573] 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).

[1574] 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.

[1575] 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.

[1576] 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).

[1577] 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.

[1578] 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.

[1579] 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.

[1580] 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.

[1581] 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.

[1582] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1583] "Example 1"

[1584] The system of the present invention has a function to send the same prompt to each AI via API as a means of sending the same prompt content to multiple AIs. Specifically, it can send the same prompt to AIs such as OpenAI via Azure and Bard via GCP.

[1585] Believe.

[1586] "Example 2"

[1587] In addition, as a means of receiving answers from multiple AIs, it has the function of receiving answers from each AI via API and managing them centrally. Specifically, it receives answers from each AI and stores them in a database.

[1588] "Example 3"

[1589] Furthermore, as a means of creating and sending new prompts based on multiple received answers, the system has the ability to analyze the received answers, generate new prompts based on them, and send them to the AI. Specifically, it extracts keywords from the received answers and generates new prompts based on them.

[1590] "Example 4"

[1591] In addition, as a means of determining the credibility of the same answer, the system has a function to compare answers received from multiple AIs and, if there are answers with the same content, rate the credibility of that answer highly. Specifically, the system compares answers received from multiple AIs and, if there are answers with the same content, rates the credibility of that answer highly.

[1592] "Example 5"

[1593] Furthermore, as a means of supporting both chat and image generation AI, it has an interface that supports both chat and image generation AI. Specifically, it has the function of receiving both text responses from chat AI and image responses from image generation AI and displaying them appropriately.

[1594] The processing flow of each embodiment will be described below.

[1595] "Example 1"

[1596] Step 1: User enters prompt.

[1597] Step 2: The system sends identical prompts to each AI via API (e.g., OpenAI via Azure, Bard via GCP).

[1598] Step 3: The system receives the answers from each AI via API.

[1599] Step 4: The system centralizes the responses received and stores them in a database.

[1600] "Example 2"

[1601] Step 1: The system retrieves the answers from each AI stored in the database.

[1602] Step 2: The system analyzes the answers obtained and extracts keywords.

[1603] Step 3: The system generates a new prompt based on the extracted keywords.

[1604] Step 4: Send the new system-generated prompts to each AI.

[1605] "Example 3"

[1606] Step 1: The system compares the answers it receives from multiple AIs.

[1607] Step 2: If there is an answer with the same content, the system will rate the credibility of that answer highly.

[1608] Step 1: The system receives both a text response from the chat AI and an image response from the image generation AI.

[1609] Step 2: The system displays the received text and image answers appropriately.

[1610] Example 1

[1611] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1612] In conventional systems, it was difficult to send the same prompt to multiple generative AI models and compare their responses. Furthermore, there was a lack of means to determine the credibility of the responses from each generative AI model, which sometimes raised doubts about the reliability of the information users received. Furthermore, there was no general-purpose system that could handle chat-based and image-based generative AI models.

[1613] 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.

[1614] In this invention, the server includes a means for a user to input a prompt text, a means for a terminal to send the prompt text to the server, a means for the server to receive the prompt text, a means for the server to send the prompt text to a generative AI model via a cloud service, a means for the server to receive responses from each generative AI model, a means for the server to return the response to the user, a means for sending the same prompt content to multiple generative AI models, a means for receiving answers from multiple generative AI models, a means for creating and sending a new prompt based on the multiple received answers, a means for determining the credibility of the same answer content, and a means for supporting chat-based and image-based generative AI models. This makes it possible to compare responses from multiple generative AI models and determine their credibility. Furthermore, a general-purpose system compatible with chat-based and image-based generative AI models can be provided.

[1615] "User" means an individual or entity that utilizes the system to enter prompts and receive responses from the generative AI model.

[1616] A "terminal" is a device on which a user inputs a prompt and sends it to a server, and includes a PC, smartphone, tablet, etc.

[1617] The "server" is a computer system whose role is to receive prompts, send them to the generative AI model, receive responses, and return them to the user.

[1618] A "prompt" is a question or instruction text that a user enters into a generative AI model.

[1619] "Cloud services" is a general term for computing resources and services provided over the internet, and is a platform that enables access to generative AI models.

[1620] A "generative AI model" is an artificial intelligence algorithm or system that generates a response to a prompt from a user.

[1621] A "response" is an answer or piece of information that a generative AI model generates based on a prompt.

[1622] "Credibility" is a criterion for assessing whether the responses from a generative AI model are accurate and trustworthy.

[1623] A "chat-based generative AI model" is a generative AI model designed for text-based dialogue.

[1624] An "image generation system generative AI model" is a generative AI model designed to generate images based on prompt text.

[1625] A "new prompt" is a new question or instruction text that is created based on multiple responses received.

[1626] This invention is a system that sends the same prompt sentence to multiple generative AI models and compares their responses. A specific embodiment of this system is described below.

[1627] The user inputs a prompt using a terminal, which can be a PC, smartphone, tablet, or other device, and sends the input prompt to the server. The server receives the prompt and sends it to the generative AI model via a cloud service.

[1628] The server uses a platform commonly used for cloud services, such as Azure or GCP, to access the generative AI model. Specifically, it uses the Azure SDK or GCP SDK to generate an API request and includes the prompt statement as the payload.

[1629] The server receives responses from each generative AI model and returns them to the user. Responses are answers or information generated by the generative AI model based on the prompt. The server compiles the received responses and returns them to the device, which then displays them to the user.

[1630] As a specific example, if a user enters a prompt such as "What's the weather like today?", the server sends this prompt to multiple generative AI models via a cloud service. For example, the server sends the same prompt to a generative AI model provided via Azure and a generative AI model provided via GCP. After receiving responses from each generative AI model, the server returns these responses to the user. The user will see the message "Generative AI model A: Today's weather is sunny. Generative AI model B: Today's weather is cloudy" displayed on their device screen.

[1631] This system allows users to compare responses from multiple generative AI models and judge their credibility. It also provides a general-purpose system that supports chat and image generation generative AI models.

[1632] Examples of prompts include:

[1633] "What's the weather like today?"

[1634] "Please tell me the latest news."

[1635] How do I write a Hello World program in Python?

[1636] In this way, users can ask a variety of questions and give instructions to the generative AI model and use the responses it receives.

[1637] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1638] Step 1:

[1639] The user enters a prompt statement.

[1640] Specifically, the user enters "What's the weather like today?" into the input field of the terminal. The entered prompt text is temporarily stored in the terminal's memory. Input: The prompt text entered by the user. Output: The prompt text stored in the terminal.

[1641] Step 2:

[1642] The terminal sends the prompt to the server.

[1643] The device uses the JavaScript fetch API to send an HTTP POST request to the server containing the prompt text. The request also includes the user's identity and a timestamp. Input: The prompt text stored on the device. Output: The HTTP request sent to the server.

[1644] Step 3:

[1645] The server receives the prompt.

[1646] The server uses the Node.js Express framework to extract the prompt text from the received HTTP request. The extracted prompt text is stored in the server's memory. Input: HTTP request sent to the server. Output: Prompt text stored on the server.

[1647] Step 4:

[1648] The server sends the prompt to the generative AI model via a cloud service.

[1649] The server uses the Azure SDK to send a prompt to the API of Generative AI Model A. At the same time, it uses the GCP SDK to send the same prompt to the API of Generative AI Model B. Input: Prompt text saved on the server. Output: Prompt text sent to Generative AI Model A and Generative AI Model B.

[1650] Step 5:

[1651] The server receives the response from each generative AI model.

[1652] The server waits for HTTP responses from Generative AI Model A and Generative AI Model B, and receives each response. The received responses are stored in the server's memory. Input: HTTP responses from Generative AI Model A and Generative AI Model B. Output: Responses stored on the server.

[1653] Step 6:

[1654] The server returns the response to the user.

[1655] The server compiles the received responses in JSON format and returns them to the device. The device then displays this to the user. Specifically, the browser screen displays "Generated AI model A: Today's weather is sunny. Generated AI model B: Today's weather is cloudy." Input: Response stored on the server. Output: Response displayed to the user.

[1656] (Application example 1)

[1657] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1658] Conventional systems using generative AI models lacked a means to efficiently compare and integrate responses from multiple AIs, making it difficult to provide information to users in an optimal format. Furthermore, the lack of consideration for implementation as an application installed on a specific device resulted in poor user convenience. This posed a challenge, limiting the scope of use of generative AI models.

[1659] 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.

[1660] In this invention, the server includes means for sending the same prompt content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending new prompts based on the received multiple answers, means for determining the credibility of the same answer content, means for supporting chat-type and image generation-type AIs, means for comparing and integrating the received multiple answers and providing them to the user in an optimal form, and means including an application to be installed on a smartphone. This makes it possible to efficiently compare and integrate answers from multiple generative AI models and provide information to the user in an optimal form.

[1661] "Artificial intelligence" refers to computer systems that have the ability to analyze data, learn, and make decisions.

[1662] A "prompt" is an input sentence that causes an artificial intelligence to perform a specific task.

[1663] An "answer" is the output generated by an artificial intelligence in response to a prompt.

[1664] "Credibility" is a criterion for evaluating the accuracy and reliability of an answer.

[1665] "Chat AI" is AI for conducting text-based conversations.

[1666] "Image generation artificial intelligence" is artificial intelligence that has the ability to generate images.

[1667] "Comparison and synthesis" is the process of evaluating multiple responses and combining them into one optimal answer.

[1668] A "smartphone" is a mobile phone that can connect to the Internet and run applications.

[1669] An "application" is a software program designed to perform a specific task.

[1670] A system for implementing this invention sends the same prompt content to multiple artificial intelligences, receives responses from the multiple artificial intelligences, compares and integrates them, and provides them to a user. The system is implemented as an application installed on a smartphone.

[1671] Hardware and software used

[1672] Hardware: Smartphone

[1673] Software: Python, cloud service APIs (Azure's OpenAI API, GCP's Bard API)

[1674] System Operation

[1675] 1. Entering a prompt: The user enters a prompt through a smartphone application, for example, "Tell me about the city of the future."

[1676] 2. Sending prompts: The server sends the input prompt to OpenAI via Azure and Bard via GCP, allowing the same prompt to be sent to multiple AIs.

[1677] 3. Receiving answers: The server receives answers from each AI. For example, OpenAI answers "The city of the future is a place where smart technology and sustainable energy solutions come together," while Bard answers "The city of the future is an environment where AI and IoT are highly integrated."

[1678] 4. Comparing and integrating answers: The server compares the multiple answers received and integrates them in the most optimal way. For example, it might integrate them as "The city of the future is an environment where smart technology and sustainable energy solutions are combined, and AI and IoT are highly integrated."

[1679] 5. Providing to users: The integrated answers are provided to users through a smartphone application, allowing them to obtain information from multiple AIs at once.

[1680] Specific examples

[1681] If the user types the prompt "Tell me about the city of the future," the system will:

[1682] Example prompt: Tell me about the city of the future.

[1683] Based on this prompt, OpenAI via Azure and Bard via GCP provide their respective answers, which are then integrated and presented to the user. For example, the integrated answer looks like this:

[1684] Example of an integrated answer: The city of the future is a blend of smart technology and sustainable energy solutions, with highly integrated AI and IoT.

[1685] In this way, users can efficiently obtain information from multiple generative AI models and utilize it in the most optimal way.

[1686] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1687] Step 1:

[1688] A user starts an application on their smartphone and inputs a prompt. The input prompt is text data such as "Tell me about the city of the future." The input prompt is then sent to the server.

[1689] Step 2:

[1690] The server sends the received prompt sentence to OpenAI via Azure and Bard via GCP. Specifically, it sends the prompt sentence using the API of each cloud service. The input is the prompt sentence, and the output is the answer from each AI.

[1691] Step 3:

[1692] The server receives answers from each AI. For example, OpenAI returns text data such as "The city of the future is a place where smart technology and sustainable energy solutions are combined," and Bard returns text data such as "The city of the future is an environment where AI and IoT are highly integrated." The input is the answer from each AI, and the output is a list of the received answers.

[1693] Step 4:

[1694] The server compares and integrates the multiple responses it receives. Specifically, it analyzes the content of each response and integrates them while eliminating duplicates. For example, it might integrate the following statement: "The city of the future is an environment where smart technology and sustainable energy solutions are combined, and AI and IoT are highly integrated." The input is a list of received responses, and the output is the integrated response.

[1695] Step 5:

[1696] The server sends the integrated answer to a smartphone application, whose input is the integrated answer and whose output is text data displayed on the user's smartphone.

[1697] Step 6:

[1698] The user can view the integrated answer through a smartphone application, which allows the user to obtain information from multiple generative AI models at once. The input is the integrated answer sent from the server, and the output is text data that the user can view.

[1699] Example 2

[1700] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1701] In conventional AI systems, it was difficult to centrally manage responses from multiple AIs and determine their credibility. Furthermore, the system lacked the ability to efficiently store received responses and search and view them later. This made it time-consuming for users to compare and consider responses from multiple AIs, hindering efficient use.

[1702] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1703] In this invention, the server includes means for sending the content of the same input sentence to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending a new input sentence based on the multiple received answers, means for determining the credibility of the same answer content, means for handling chat-type and image-generating AIs, means for centrally managing answers from each AI, means for saving received answers in a database, and means for searching and viewing saved answers. This allows answers from multiple AIs to be efficiently managed, their credibility to be determined, and users to easily search and view them.

[1704] "Artificial intelligence" refers to computer systems that have the ability to learn from data and perform specific tasks automatically.

[1705] An "input sentence" is a textual instruction or question given by a user to an artificial intelligence.

[1706] An "answer" is a text or other form of output that an artificial intelligence generates based on an input sentence.

[1707] "Credibility" is a criterion for assessing whether answers from artificial intelligence are accurate and trustworthy.

[1708] "Chat AI" is AI designed for text-based conversations.

[1709] "Image generation artificial intelligence" is artificial intelligence that has the ability to generate images based on input text and other data.

[1710] "Centralized management" means integrating and managing multiple pieces of information and data in one system or location.

[1711] A "database" is a system for efficiently storing, searching, and managing data.

[1712] "Searching and viewing" means searching for stored data based on specific conditions and checking its contents.

[1713] The present invention relates to a system for centrally managing responses from multiple artificial intelligences and determining their credibility. Specific embodiments of this system will be described below.

[1714] The server sends the prompt text entered by the user to multiple generative AI models. Generative AI models include models specialized in text generation and image generation. Specific examples of generative AI models include OpenAI's GPT-3 and Google's BERT.

[1715] The server receives the prompt text entered by the user and sends an API request to each generative AI model. The API request is made using an HTTP POST request, sending data including the prompt text. The answer from each generative AI model is returned in JSON format.

[1716] The server receives the answers from each generative AI model and stores them in a database, typically a relational database such as MySQL or PostgreSQL. The stored data includes prompt statements and the corresponding answers from each generative AI model.

[1717] Users can search and view past prompts and their responses through the system interface. For example, they can display a list of responses from each generative AI model to the prompt, "Tell me about the future of AI-based text generation."

[1718] For example, if a user enters the prompt "Please tell me about the future of AI text generation," the server sends this prompt to each generative AI model. The server receives the answers from each generative AI model and stores them in a database. The user can later compare the answers of each generative AI model to this prompt.

[1719] Examples of prompts include:

[1720] "Please tell me about the future of AI text generation."

[1721] "Please explain the effects of climate change."

[1722] "Tell me about the latest technology trends"

[1723] This system efficiently manages answers from multiple generative AI models, determines their credibility, and allows users to easily search and view them.

[1724] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1725] Step 1:

[1726] The user enters a prompt statement.

[1727] The user enters a prompt through the system interface, for example, by typing "Tell me about the future of AI-generated text" into a text box on a web browser and clicking the submit button.

[1728] Input: The prompt text entered by the user.

[1729] Output: The prompt text sent to the server.

[1730] Step 2:

[1731] The server sends the prompt sentence to each generative AI model.

[1732] The server receives the prompt sentence entered by the user and sends an API request to each generative AI model. Specifically, it sends the prompt sentence to the API endpoint of the generative AI model using an HTTP POST request.

[1733] Input: The prompt received from the user.

[1734] Output: API requests sent to each generative AI model.

[1735] Step 3:

[1736] The server receives the answers from each generative AI model.

[1737] The server receives API responses from each generative AI model. Responses are usually returned in JSON format. For example, from GPT-3, the server receives the answer, "The future of AI text generation is very bright."

[1738] Input: API response from each generative AI model.

[1739] Output: The response data received by the server.

[1740] Step 4:

[1741] The server stores the received response in a database.

[1742] The server saves the received answers in a database. Specifically, it uses a relational database such as MySQL or PostgreSQL to store the prompt sentences and the corresponding answers of each generative AI model in a table.

[1743] Input: The answer data received from each generative AI model.

[1744] Output: Response data stored in a database.

[1745] Step 5:

[1746] Users can search and view saved answers.

[1747] Users can search and view past prompts and their responses through the system interface. For example, the system displays a list of responses from each generative AI model to the prompt, "Tell me about the future of AI-based text generation."

[1748] Input: The user's search query.

[1749] Output: The prompt and answer data displayed as search results.

[1750] (Application example 2)

[1751] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1752] Conventional AI systems have difficulty managing responses from multiple AIs in a unified manner and providing users with the most appropriate answer. Furthermore, they lacked the means to determine the credibility of the answers, making it impossible to provide users with highly reliable information. Furthermore, they lacked the ability to save answers in a database and refer to them later, making efficient information management impossible.

[1753] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1754] In this invention, the server includes means for sending the same input content to multiple AIs, means for receiving answers from the multiple AIs, means for creating and sending new input content based on the received multiple answers, means for determining the credibility of the same answer content, means for handling chat-type and image-generating AIs, means for saving answers from the multiple AIs in a database, means for selecting the optimal answer from the saved answers, and means for providing the selected optimal answer to the user. This makes it possible to centrally manage answers from the multiple AIs and provide the user with a highly reliable optimal answer.

[1755] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and learn, reason, and self-correct.

[1756] "Input" refers to questions or instructions provided by users to the artificial intelligence.

[1757] An "answer" is a response or information generated by artificial intelligence based on input.

[1758] A "database" is a system for efficiently storing, managing, and retrieving information.

[1759] "Credibility" is an indicator of whether information or answers are accurate and trustworthy.

[1760] "Chat AI" is AI that can conduct text-based conversations.

[1761] "Image generation artificial intelligence" is artificial intelligence that can generate images based on input content.

[1762] The "best answer" is the most appropriate and reliable answer selected from multiple answers.

[1763] "User" refers to a person or organization that uses the system to ask questions or give instructions.

[1764] "Cloud services" refer to computer resources and services provided over the Internet.

[1765] A "network" is a system in which multiple computers and devices are connected to each other and used to exchange information.

[1766] The system for carrying out this invention has the function of sending the same input content to multiple AIs, receiving responses from the multiple AIs, and managing them in a unified manner. A specific embodiment of this system will be described below.

[1767] System Configuration

[1768] The server includes the following means:

[1769] 1. Input content transmission means: Sends questions and instructions (prompt text) entered by the user to multiple artificial intelligences.

[1770] 2. Answer receiving means: Receive answers from multiple artificial intelligences.

[1771] 3. Method for creating new input content: Create new input content based on the multiple answers received and send it to the artificial intelligence again.

[1772] 4. Credibility judgment method: judge the credibility of the same answers.

[1773] 5. Database storage means: Responses from multiple artificial intelligences are stored in a database.

[1774] 6. Best answer selection method: Select the best answer from the saved answers.

[1775] 7. User-provided means: The selected best answer is provided to the user.

[1776] Hardware and software used

[1777] Hardware: Servers and user devices (smartphones, PCs, etc.) that can connect to the Internet.

[1778] Software: Python, SQLite, APIs for various artificial intelligence models (e.g., GPT-4, BERT, T5, etc.).

[1779] Data processing and calculation

[1780] The server receives input from the user and sends it to multiple AIs. After receiving answers from each AI, the server stores these answers in a database. The server selects the best answer from the stored answers and provides it to the user. The credibility judgment means evaluates the credibility of the same answer content and provides highly reliable information.

[1781] Specific examples

[1782] When a user asks, "What's the weather like today?", the server sends this prompt to multiple AIs. For example, GPT-4 may give an answer like, "It's sunny today," BERT may give an answer like, "It's cloudy today," and T5 may give an answer like, "It's going to rain today." These answers are stored in a database and evaluated using a credibility assessment tool. The best answer is selected and served to the user.

[1783] Example prompt sentence:

[1784] What's the weather like today?

[1785] In this way, it becomes possible to centrally manage answers from multiple AIs and provide users with the most reliable and optimal answers.

[1786] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1787] Step 1:

[1788] The user uses the terminal to enter input (prompt text).

[1789] Input: The prompt text entered by the user (e.g., "What's the weather like today?").

[1790] Output: The input is sent to the server.

[1791] Specific operation: When the user enters a question in the input field of the terminal and presses the send button, the prompt sentence is sent to the server.

[1792] Step 2:

[1793] The server receives the input and sends it to multiple generative AI models.

[1794] Input: The prompt text sent by the user.

[1795] Output: The prompt sentence is sent to multiple generative AI models.

[1796] Specific operation: The server sends the received prompt sentence to the API endpoint of each generative AI model (GPT-4, BERT, T5, etc.).

[1797] Step 3:

[1798] The server receives the answers from each generative AI model.

[1799] Input: The prompt sent to each generative AI model.

[1800] Output: The answer from each generative AI model.

[1801] Specific operation: Each generative AI model generates an answer based on the prompt sentence and sends the answer back to the server.

[1802] Step 4:

[1803] The server stores the received response in a database.

[1804] Input: The answers from each generative AI model.

[1805] Output: The answer is saved in the database.

[1806] Specific operation: The server stores the received answer in an SQLite database, along with information to identify which generative AI model the answer came from.

[1807] Step 5:

[1808] The server selects the best answer from the stored answers.

[1809] Input: Multiple answers stored in a database.

[1810] Output: The best answer.

[1811] What it does: The server analyzes the stored answers and runs an algorithm to select the best answer based on the authenticity and content of the answer.

[1812] Step 6:

[1813] The server provides the best answer to the user.

[1814] Input: Best answer.

[1815] Output: The answer provided to the user.

[1816] Specific operation: The server sends the selected optimal answer to the user's device and displays it on the user's screen.

[1817] In this way, it is possible to centrally manage responses from multiple generative AI models to prompts entered by users, and provide highly reliable and optimal answers.

[1818] Example 3

[1819] Next, a third embodiment of the third embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1820] Conventional systems have difficulty effectively analyzing responses from multiple generative AI models and generating new prompts. Furthermore, they lack a means to determine the credibility of responses from multiple generative AI models, making it impossible to provide users with reliable information. Furthermore, there was no unified interface for chat and image generation generative AI models, resulting in an inconsistent user experience.

[1821] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1822] In this invention, the server includes means for receiving input from a user, means for sending prompts to multiple generative AI models, means for receiving responses from the multiple generative AI models, means for analyzing the received responses and extracting keywords, means for generating new prompts based on the extracted keywords, means for comparing responses from the multiple generative AI models and determining their credibility, and means for supporting chat-based and image-based generative AI models. This makes it possible to effectively analyze responses from multiple generative AI models based on user input and generate new prompts. Furthermore, by determining the credibility of responses from multiple generative AI models, highly reliable information can be provided to the user. Furthermore, by providing a unified interface for chat-based and image-based generative AI models, a consistent user experience can be achieved.

[1823] A "user" is an entity that inputs prompts into the system and receives answers from the generative AI model.

[1824] A "server" is a device or system that receives input from users, sends prompts to multiple generative AI models, and receives and analyzes responses.

[1825] A "generative AI model" is an artificial intelligence model that generates answers such as text and images based on a given prompt.

[1826] A "prompt" is text that contains instructions or questions that are input to a generative AI model.

[1827] An "answer" is an output result, such as text or an image, that a generative AI model generates based on a prompt.

[1828] "Keywords" are important words or phrases extracted by analyzing the answers from the generative AI model.

[1829] "Credibility" is the reliability of an answer, assessed based on the degree to which answers from multiple generative AI models match.

[1830] A "chat-based generative AI model" is a generative AI model for conducting text-based conversations.

[1831] An "image generation system generative AI model" is a generative AI model for generating images based on prompt sentences.

[1832] An "interface" is a means or method by which a user interacts with a system, and corresponds to the generative AI models of chat and image generation systems.

[1833] MODE FOR CARRYING OUT THE INVENTION

[1834] This invention relates to a system that receives input from a user, analyzes responses from multiple generative AI models, and generates and sends new prompt sentences. A specific embodiment of this system is described below.

[1835] Hardware and software used

[1836] The server uses the following hardware and software:

[1837] Hardware: A server machine with a powerful processor, sufficient memory, and a fast network connection.

[1838] Software: Operating systems (e.g., Linux), programming languages ​​(e.g., Python), natural language processing libraries (e.g., spaCy, NLTK), text matching algorithms (e.g., difflib, fuzzywuzzy), generative AI models (e.g., OpenAI's GPT-3, DALL-E).

[1839] System processing flow

[1840] A user uses a terminal to input a prompt to the system, such as "Generate an image of a cat" or "What is the diameter of the Earth?"

[1841] The server receives input from the user and sends prompts to multiple generative AI models using the APIs of the generative AI models (e.g., OpenAI's ChatGPT or DALL-E).

[1842] The server receives answers from each generative AI model, which may be in text or image format.

[1843] The server analyzes the received responses and extracts keywords using Python's natural language processing libraries, such as spaCy and NLTK. For example, in response to a prompt such as "Generate an image of a cat," the server extracts keywords such as "cat" and "image."

[1844] The server generates a new prompt based on the extracted keywords. This prompt is generated using a generative AI model (e.g., OpenAI's GPT-3). For example, it generates a new prompt such as "Generate a more detailed image of a cat."

[1845] The server compares answers from multiple generative AI models and, if there are identical answers, rates them as more reliable. This comparison uses text matching algorithms like "difflib" and "fuzzywuzzy."

[1846] The server sends the most credible answers to the terminal for display to the user, and the terminal shows the displayed text answers and image answers to the user.

[1847] Specific examples

[1848] Example 1: Receiving input from the user

[1849] A user types "generate an image of a cat" into a terminal.

[1850] The server receives this input and logs it.

[1851] Example 2: Sending prompts to multiple AIs

[1852] The server sends a prompt to ChatGPT and DALL-E APIs: "Generate a cat image."

[1853] The server checks the status of the API request and proceeds to the next step if successful.

[1854] Example 3: Receiving answers from AI

[1855] The server receives text responses from ChatGPT and image responses from DALL-E.

[1856] The server temporarily stores the received response.

[1857] Example 4: Answer analysis and keyword extraction

[1858] The server analyzes ChatGPT's text responses using "spaCy" and extracts keywords such as "cat" and "image."

[1859] The server records the extracted keywords in a log.

[1860] Example 5: Creating a new prompt

[1861] Based on the extracted keywords, the server generates a new prompt such as "Generate a more detailed image of a cat."

[1862] The server sends this prompt to the generative AI model.

[1863] Example 6: Judging credibility

[1864] The server uses "fuzzywuzzy" to compare answers received from multiple generative AI models, and if there are answers with the same content, it rates their credibility highly.

[1865] The server logs the results of the credibility assessment.

[1866] Example 7: Displaying results to the user

[1867] The server sends the most credible answer to the terminal for display to the user.

[1868] The terminal shows the displayed text and image answers to the user.

[1869] In this way, this system can effectively analyze responses from multiple generative AI models based on user input and generate new prompts. Furthermore, by determining the credibility of responses from multiple generative AI models, it is possible to provide highly reliable information to the user. Furthermore, by providing a unified interface for chat-based and image-based generative AI models, it is possible to ensure a consistent user experience. The flow of the identification process in Example 3 is described with reference to FIG. 15.

[1870] Step 1:

[1871] A user uses a terminal to input a prompt to the system, such as "Generate an image of a cat" or "What is the diameter of the Earth?"

[1872] Input: The prompt text entered by the user.

[1873] Output: The prompt received by the server.

[1874] Specific behavior: The user enters a prompt into the input field of the terminal and presses the send button. The server receives this input and records it in the log.

[1875] Step 2:

[1876] The server sends the received prompts to multiple generative AI models using the APIs of the generative AI models (e.g., OpenAI's ChatGPT or DALL-E).

[1877] Input: The prompt received by the server.

[1878] Output: The prompt sent to each generative AI model.

[1879] Specific operation: The server sends the prompt sentence as an API request to the generative AI model, checks the status of the API request, and proceeds to the next step if it is successful.

[1880] Step 3:

[1881] The server receives answers from each generative AI model, which may be in text or image format.

[1882] Input: The answer from each generative AI model.

[1883] Output: The answer stored on the server.

[1884] Specific operation: The server receives the API response from the generative AI model and temporarily stores the text and image answers.

[1885] Step 4:

[1886] The server analyzes the received responses and extracts keywords using Python's natural language processing libraries, spaCy and NLTK.

[1887] Input: The received text response.

[1888] Output: Extracted keywords.

[1889] Specific operation: The server analyzes the text response using "spaCy" and "NLTK" to extract keywords such as "cat" and "image." The extracted keywords are recorded in a log.

[1890] Step 5:

[1891] The server generates a new prompt based on the extracted keywords, using a generative AI model (e.g., OpenAI's GPT-3).

[1892] Input: Extracted keywords.

[1893] Output: The new prompt text.

[1894] Specific operation: The server generates a new prompt sentence based on the extracted keywords, such as "Generate a more detailed image of a cat," and sends this prompt sentence to the generative AI model.

[1895] Step 6:

[1896] The server compares answers from multiple generative AI models and, if there are identical answers, rates them as more reliable. This comparison uses text matching algorithms like "difflib" and "fuzzywuzzy."

[1897] Input: Answers from multiple generative AI models.

[1898] Output: The credibility assessment result.

[1899] Specific behavior: The server compares the received answers using "difflib" or "fuzzywuzzy", and if there are answers with the same content, it rates the answer as more reliable. It records the result of the reliability evaluation in a log.

[1900] Step 7:

[1901] The server sends the most credible answers to the terminal for display to the user, and the terminal shows the displayed text answers and image answers to the user.

[1902] Input: A credible answer.

[1903] Output: The answer that is displayed to the user.

[1904] Specific operation: The server sends a credible answer to the device, which displays it to the user. The user checks the displayed text and image answers.

[1905] (Application example 3)

[1906] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[1907] Conventional advertising material generation systems make it difficult for advertising creators to efficiently generate high-quality advertising copy and images. Furthermore, there was a lack of means to evaluate the credibility of responses from multiple AIs, which led to problems with the quality and reliability of the generated advertising materials. Furthermore, there was no interface compatible with both conversational AI and image-generating AI, making the advertising material generation process cumbersome.

[1908] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means. In this invention, the server includes means for sending the same instruction content to multiple AIs, means for receiving answers from multiple AIs, means for creating and sending new instructions based on the multiple received answers, means for determining the credibility of the same answer content, means for supporting dialogue-based AI and image-generating AI, means for generating instructions for generating advertising material, means for acquiring advertising copy and images from multiple generative AI models, means for evaluating the credibility of the acquired advertising copy and images, and means for displaying the acquired advertising copy and images. This enables advertising creators to efficiently generate high-quality, reliable advertising material.

[1909] "Multiple artificial intelligences" refers to multiple AI systems that operate using different algorithms and platforms.

[1910] "Instructions" refer to input information or commands that instruct the artificial intelligence to perform a specific task.

[1911] An "answer" is the output result generated by the artificial intelligence based on the instructions.

[1912] "Credibility" is the standard by which we evaluate whether the answers we receive are accurate and trustworthy.

[1913] "Conversational artificial intelligence" refers to AI systems that have the ability to converse with humans through text or voice.

[1914] "Image-generating artificial intelligence" refers to an AI system that has the ability to generate images based on input instructions.

[1915] "Advertising Materials" means content such as text and images used in an advertising campaign.

[1916] A "generative artificial intelligence model" is an AI model that generates content such as text and images for a specific task.

[1917] "Measures for assessing credibility" refers to methods or algorithms for assessing the accuracy and reliability of answers received from multiple AIs.

[1918] "Displaying means" refers to a method or device for visually presenting the generated advertising copy and images to the user.

[1919] The system for implementing this invention is a tool that allows advertising creators to efficiently generate high-quality advertising materials. The system has the function of sending the same instructions to multiple AIs, receiving responses from the multiple AIs, and creating and sending new instructions based on those responses. It also judges the credibility of the same responses and has an interface that supports dialogue-based AI and image-generating AI.

[1920] Hardware and software used

[1921] Hardware:

[1922] Smartphone

[1923] software:

[1924] Python

[1925] OpenAI API

[1926] PIL (Python Imaging Library)

[1927] requests

[1928] Data processing and calculation

[1929] Server Action:

[1930] 1. Prompt generation: Based on the keywords and themes entered by the ad creator, a prompt is generated to be sent to the generative AI model. For example, if the keyword "new smartphone release" is entered, the prompt will be as follows:

[1931] Ad theme: New smartphone launch. Generate the best copy and images.

[1932] 2. Obtaining answers from AI: Using OpenAI's API, we obtain ad copy and images from the generative AI model. Text answers are obtained from the conversational AI, and image answers are obtained from the image-generating AI.

[1933] 3. Credibility assessment: Compare the responses received to see if any of them are identical, thereby assessing their credibility.

[1934] 4. Displaying the results: Displaying the generated ad copy and fetching and displaying images. For example, if the generated ad copy is "This new smartphone is equipped with the latest technology" and the generated image is a high-resolution image of a smartphone, they will be visually presented to the ad creator.

[1935] Specific examples

[1936] When an ad creator enters the keyword "new smartphone release," the system generates the following prompt and sends it to the AI:

[1937] Ad theme: New smartphone launch. Generate the best copy and images.

[1938] The AI ​​generates the ad copy, "This new smartphone is equipped with the latest technology," along with a high-resolution image of the smartphone. The system evaluates these responses and, if deemed credible, displays them to the ad creator.

[1939] In this way, ad creators can efficiently generate high-quality, reliable ad materials.

[1940] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1941] Step 1:

[1942] The user inputs keywords or themes for generating advertising materials. For example, the input keyword might be "new smartphone release." This input becomes the basis for generating prompts.

[1943] Step 2:

[1944] The server generates a prompt based on the input keywords. Specifically, it creates the prompt "Advertisement theme: New smartphone release. Please generate the best ad copy and image." This prompt is used in the next step.

[1945] Step 3:

[1946] The server sends the generated prompt sentence to multiple generative AI models. In this case, the prompt sentence is sent to both conversational AI and image generation AI. The input is the prompt sentence, and the output is the answer from each AI model.

[1947] Step 4:

[1948] The server receives responses from multiple generative AI models. The conversational AI model returns ad copy, while the image-generating AI model returns ad images. For example, a high-resolution image of a smartphone might be generated along with the ad copy, "This new smartphone is equipped with the latest technology."

[1949] Step 5:

[1950] The server evaluates the credibility of the received answers. Specifically, it compares answers from multiple AIs to see if there are any answers with the same content. Answers that are judged to be highly credible are used in the next step.

[1951] Step 6:

[1952] The server displays the advertisement copy and images that are judged to be highly credible to the user. The user can review the generated advertisement copy and images and make corrections or adopt them as necessary. The output is the advertisement material presented to the user.

[1953] In this way, users can efficiently generate high-quality and reliable advertising materials.

[1954] 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.

[1955] "Example 1"

[1956] One embodiment of the present invention is a system incorporating an emotion engine. This system analyzes emotions from text and voice input from a user. Specifically, if a user inputs the text "Today was fun" to the system, the emotion engine recognizes the emotion "joy" from this text. Also, if a user inputs the text "Today was hard," the emotion engine recognizes the emotion "sadness."

[1957] "Example 2"

[1958] Furthermore, the emotion engine adjusts the AI's response based on the analyzed emotions. For example, if a user enters the text "I had fun today," and the emotion "joy" is recognized, the AI ​​will respond positively, such as "That's great! What was fun about it?" On the other hand, if a user enters the text "I had a hard day today," and the emotion "sadness" is recognized, the AI ​​will respond with an empathetic response, such as "That must have been tough. Maybe talking about it will help you feel a little better."

[1959] "Example 3"

[1960] In this way, a system incorporating an emotion engine can properly recognize the user's emotions and generate AI responses accordingly, enabling more human-like communication, which makes the user's interaction with the system feel more natural and satisfying.

[1961] The processing flow of each embodiment will be described below.

[1962] "Example 1"

[1963] Step 1: The user inputs text into the system. For example, the user inputs the text "I had a great time today."

[1964] Step 2: The emotion engine analyzes the emotion from the user's text input. In this example, the emotion "joy" is recognized.

[1965] "Example 2"

[1966] Step 1: The user inputs text into the system, for example, "Today was a tough day."

[1967] Step 2: The emotion engine analyzes the emotion from the user's text input. In this example, the emotion "sadness" is recognized.

[1968] Step 3: The emotion engine adjusts the AI's response based on the analyzed emotions. In this example, it generates an empathetic response such as, "That must be tough. Maybe talking about it will help you feel a little better."

[1969] "Example 3"

[1970] Step 1: The user provides text input to the system.

[1971] Step 2: The emotion engine analyzes the emotion from the user's text input.

[1972] Step 3: The emotion engine adjusts the AI's response based on the analyzed emotions.

[1973] Step 4: The AI-generated response is returned to the user, making the user's interaction with the system feel more natural and satisfying.

[1974] Example 1

[1975] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1976] In conventional systems, it was difficult to send the same prompt to multiple AIs and collect their responses. Furthermore, the system lacked the ability to analyze the user's emotions and generate responses based on them, making it impossible to provide appropriate responses that met the user's needs. Furthermore, there was also an insufficient means to determine the credibility of responses from multiple AIs.

[1977] 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.

[1978] In this invention, the server includes means for sending the same prompt content to multiple AIs, means for receiving answers from the multiple AIs, means for creating and sending new prompts based on the received multiple answers, means for determining the credibility of the same answer content, means for analyzing emotions from user input, means for generating responses based on the analyzed emotions, and means for supporting chat-type and image-generating AIs. This makes it possible to efficiently collect responses from the multiple AIs and provide appropriate responses based on the user's emotions.

[1979] "Artificial intelligence" is the technology that enables computer systems to learn, reason, and self-correct by imitating human intelligence.

[1980] A "prompt" is an input sentence that causes an artificial intelligence to generate a particular response.

[1981] "Emotion analysis" is a technology that identifies emotions from user input and evaluates the type and intensity of those emotions.

[1982] "Credibility" is the standard by which information or responses are judged to be accurate and trustworthy.

[1983] "Chat AI" is AI designed for text-based conversations.

[1984] "Image generation AI" is AI designed to generate images based on input data.

[1985] A "platform" is the underlying hardware and software environment for running software and services.

[1986] A "network" is a system that allows multiple computers and devices to communicate with each other.

[1987] "Merchandise" refers to goods and services that are the subject of commercial activity.

[1988] This invention is a system that sends the same prompt content to multiple artificial intelligences, collects their responses, analyzes the user's emotions, and generates an appropriate response. Specific embodiments of this system will be described below.

[1989] The server is connected to a terminal for receiving prompt sentences from the user. When the user inputs a prompt sentence using the terminal, the prompt sentence is sent to the server. The server passes the received prompt sentence to an emotion engine, which analyzes the user's emotion. The emotion engine identifies emotions from the user's input text and evaluates the type and intensity of the emotion.

[1990] The server then sends prompts to multiple AIs based on the analyzed emotional information. Specifically, the server uses the API of the cloud platform to send the same prompt to different AIs. For example, the server uses the API of cloud service A to send a prompt to AI model A, and simultaneously sends the same prompt to AI model B using the API of cloud service B.

[1991] Each AI model generates a response to the received prompt. The generated response is sent to the server. The server receives responses from multiple AI models and judges the credibility of those responses. The credibility judgment takes into account the degree of similarity of the response content and past response history.

[1992] Finally, the server returns a highly credible response to the user. For example, if a user inputs a prompt sentence such as "What's the weather like today?", the server receives a response such as "It's sunny today" from AI model A of cloud service A, and a response such as "It's cloudy today" from AI model B of cloud service B. The server compares these responses and returns a highly credible response to the user.

[1993] As a concrete example, the processing when the user inputs the prompt sentence "What's the weather like today?" is shown below.

[1994] 1. A user uses a terminal to type, "What's the weather like today?"

[1995] 2. The server receives the prompt and performs emotion analysis using the emotion engine.

[1996] 3. The server uses the API of cloud service A to send a prompt to artificial intelligence model A, and simultaneously sends the same prompt to artificial intelligence model B using the API of cloud service B.

[1997] 4. AI model A responds, "It's sunny today," and AI model B responds, "It's cloudy today."

[1998] 5. The server receives these responses and determines their authenticity.

[1999] 6. The server returns a reliable response to the user.

[2000] In this way, the server can efficiently collect responses from multiple artificial intelligences and provide an appropriate response based on the user's emotions.

[2001] The flow of the identification process in the first embodiment will be described with reference to FIG.

[2002] Step 1:

[2003] The user enters a prompt statement.

[2004] The user uses a terminal to input a prompt to the system, for example, "What's the weather like today?" The input prompt is sent from the terminal to the server.

[2005] Input: User prompt text

[2006] Output: The prompt sent to the server

[2007] Step 2:

[2008] The server receives the prompt.

[2009] The server receives the prompt sent by the user, which becomes the input data for the next processing step.

[2010] Input: Prompt sent from the terminal

[2011] Output: Received prompt text

[2012] Step 3:

[2013] The server performs emotion analysis using an emotion engine.

[2014] The server passes the received prompt sentence to the emotion engine for emotion analysis. For example, if the user enters "I had fun today," the emotion engine recognizes the emotion of "joy."

[2015] Input: Received prompt text

[2016] Output: Parsed emotion information

[2017] Step 4:

[2018] The server sends prompt sentences to multiple artificial intelligences.

[2019] The server uses the API of the cloud platform to send the same prompt to different AIs, for example...

Claims

1. A means for acquiring a prompt sentence transmitted from a terminal device; means for transmitting the prompt sentence to a plurality of different generative AI models; A means for obtaining a response generated by each of the plurality of generative AI models based on the prompt sentence; means for transmitting a response by each of the plurality of generative AI models to the terminal device; A system including:

2. A means for acquiring a prompt sentence transmitted from a terminal device; means for transmitting the prompt sentence to a plurality of different generative AI models; A means for obtaining a response generated by each of the plurality of generative AI models based on the prompt sentence; A means for evaluating the credibility of responses including the same content among the responses by each of the plurality of generative AI models and transmitting the most credible response to the terminal device; system.

3. A means for acquiring a prompt sentence transmitted from a terminal device; means for transmitting the prompt sentence to a plurality of different generative AI models; A means for obtaining a response generated by each of the plurality of generative AI models based on the prompt sentence; means for aggregating responses from each of the plurality of generative AI models; means for transmitting a consolidated response to said terminal device; A system including:

Images