Information processing system, computer, information processing method and program

Abstract

Achieve both cost and safety. [Solution] An information processing system that utilizes a generation AI that generates answers to prompts acquires a prompt, determines whether an on-premise generation AI or a cloud-based generation AI will be the destination generation AI for the prompt, inputs the prompt into the destination generation AI based on the determination result, acquires an answer to the prompt from the destination generation AI that input the prompt, and outputs the answer.

Description

[Technical Field]

[0001] The present invention relates to a technology that is effective in utilizing generative AI (Artificial Intelligence). [Background technology]

[0002] Generative AI has become increasingly popular in recent years. For example, Patent Document 1 discloses a technology in which a script including target data and commands is input to a generation AI, and output data from the generation AI is stored in a database and presented to a user. Additionally, Patent Document 2 discloses a technology for evaluating the financial risk of a target company using a text generation AI. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2025-13259 [Patent Document 2] Patent No. 7626411 Summary of the Invention [Problem to be solved by the invention]

[0004] Even when users want to use generative AI, there are also a certain number of people who want to avoid having their data uploaded to the cloud, even if only temporarily. One possible solution for such users would be to provide generative AI that is always available in a closed network environment (on-premise, in-house LAN (Local Area Network), VPN (Virtual Private Network), etc.), but this would require enormous resources and would not be realistic. Therefore, there is a need for technology that can achieve both cost and security by using a closed network environment only when the user uses the generating AI with information that the user wants to keep confidential, and by using an open network environment (such as the Internet, Wi-Fi, or SaaS (Software as a Service) services) when the user uses the generating AI with other information. However, the techniques described in Patent Documents 1 and 2 do not have a mechanism for using different network environments depending on whether the information is to be kept secret or not.

[0005] Therefore, the inventor focused on a mechanism that appropriately determines whether a closed or open network environment is required in response to a prompt, and obtains an answer from the generation AI using the network environment based on the determination result.

[0006] In view of the above problems, an object of the present invention is to provide an information processing system, a computer, an information processing method, and a program that can achieve both cost reduction and safety. [Means for solving the problem]

[0007] The present invention provides an information processing system that utilizes a generation AI to generate answers to prompts, No. 1 a prompt acquisition unit that acquires a prompt; The aforementioned No. 1 The prompt will ask you to select whether the connected generation AI will be an on-premise generation AI or a cloud-based generation AI. a prompt generator that generates a second prompt for determining whether or not the first prompt includes confidential information; Based on the answer to the second prompt, the first prompt determines whether the on-premise generation AI or the cloud-based generation AI is to be the destination generation AI. a connection destination determination unit for determining The connection destination generation AI based on the judgment result No. 1 a prompt input section for inputting a prompt; The aforementioned No. 1 From the connection generation AI that entered the prompt, No. 1 an answer acquisition unit that acquires answers to the prompts; a response output unit that outputs the response; An information processing system is provided.

[0008] According to the present invention, a determination is made in response to a prompt as to whether an on-premise generation AI or a cloud-based generation AI will be the destination generation AI, and an answer to the prompt is obtained from the destination generation AI based on the determination result. As a result, the system determines whether the network environment is closed or open depending on the prompt, and security can be maintained without requiring enormous resources, making it possible to achieve both cost and security.

[0009] Although the present invention is in the category of a system, the same functions and effects can be achieved in other categories such as a computer, a method, and a program. [Effects of the Invention]

[0010] According to the present invention, it is possible to achieve both cost reduction and safety. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a diagram illustrating an overview of an information processing system 1. [Figure 2] FIG. 1 is a diagram illustrating a functional configuration of an information processing system 1. [Figure 3] FIG. 10 is a flowchart showing a response process executed by the computer 10. [Figure 4] FIG. 2 is a diagram schematically illustrating a prompt acquired by the computer 10. [Figure 5] FIG. 10 is a flowchart of a first connection destination determination process executed by the computer 10. [Figure 6] FIG. 1 is a diagram illustrating a classification prompt generated by computer 10. [Figure 7] FIG. 10 is a flowchart of a second connection destination determination process executed by the computer 10. [Figure 8] FIG. 2 is a diagram illustrating a decision prompt generated by the computer 10. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, embodiments of the present invention (hereinafter referred to as "embodiments") will be described in detail with reference to the accompanying drawings. In the following drawings, the same elements are designated by the same numbers or symbols throughout the description of the embodiments.

[0013] In this specification, "generative AI" generally refers to generative artificial intelligence technology that generates new data or content (data not limited to text characters but including images, videos, programs, etc.) based on given input data, and may be realized by a chatbot that utilizes a large-scale language model.

[0014] [Outline of Information Processing System 1] 1 is a schematic diagram for explaining an overview of an information processing system 1. Components of the information processing system 1 will be described based on FIG. The information processing system 1 is a system that includes at least a computer 10 that has a server function and utilizes a generation AI that generates answers to prompts. In this embodiment, in addition to the computer 10, the information processing system 1 also includes a user terminal 3 used by a user who asks questions to the generation AI and receives answers, an on-premise computer 4 that provides an on-premise generation AI (a generation AI provided in a closed network environment), and a cloud computer 5 that provides a cloud-based generation AI (a generation AI provided in an open network environment).

[0015] The user terminal 3 is a terminal device used by a user, such as a mobile phone, smartphone, tablet terminal, personal computer, laptop computer, or a wearable terminal such as a smart watch, smart glasses, or HMD (Head Mounted Display).

[0016] The on-premise computer 4 is a computer that exists in a closed network environment and has server functionality. The on-premise computer 4 stores vector data and file summaries of confidential information of companies, organizations and individuals, such as important information (customer information, personnel information, financial information, business plans, product technology information, system information, etc.), confidential information (minutes, customer lists, proposals, contracts, personnel information, accounting information, etc.), and personal information (name, gender, age, address, telephone number, job title, wages, family composition, etc.). The on-premise computer 4 also has on-premise generation AI functionality. There is only one on-premise computer 4 in the system.

[0017] The cloud computer 5 is a computer that exists in an open network environment and has a server function. The cloud computer 5 has the function of a cloud-based generation AI such as a SaaS-based generation AI.

[0018] The computer 10 has a server function and may be realized, for example, by a single computer, or may be realized by multiple computers like a cloud computer. In this specification, a cloud computer may refer to either a computer that uses any computer in a scalable manner to perform a specific function, or a computer that includes multiple functional modules to realize a system and uses the functions in any combination. In addition, the information processing system 1 may include other terminals and devices in addition to the above-mentioned user terminal 3, on-premise computer 4, cloud computer 5, and computer 10, and the number, type, and functions thereof are not particularly limited and can be designed as appropriate.

[0019] The present invention uses generative AI (such as a large-scale language model that performs natural language processing). The generative AI may be any AI that outputs a response to a predetermined input prompt. In the information processing system 1, the computer 10 has the function of generating AI, and the on-premise computer 4 and the cloud computer 5 also have the function of generating AI.

[0020] When computer 10 uses the function of the generation AI that it possesses, it inputs a predetermined prompt to the generation AI and obtains the answer output by the generation AI. Also, when computer 10 uses the function of the generation AI that on-premise computer 4 or cloud computer 5 possesses, it outputs a prompt to on-premise computer 4 or cloud computer 5, and obtains the answer obtained as a result of on-premise computer 4 or cloud computer 5 inputting this prompt to the generation AI from on-premise computer 4 or cloud computer 5, etc., and obtains the answer to this prompt.

[0021] The contents of the generative AI itself, such as the data that the generative AI learns, the learning method, and the method for constructing the learning model, are not particularly limited. Furthermore, the generative AI may be general-purpose or may be tuned according to the intended use and method of use, and is not particularly limited. In addition to the above reasons, since the present invention is an invention for utilizing the generative AI, a detailed description of the generative AI will be omitted to simplify the description of the specification.

[0022] An overview of the processing steps when the information processing system 1 utilizes a generation AI that generates an answer to a prompt will be described.

[0023] The computer 10 receives a prompt (step S1). The computer 10 acquires from the user terminal 3 the prompt input by the user.

[0024] The computer 10 determines whether the prompt should select an on-premise generation AI or a cloud-based generation AI as the connection destination generation AI (step S2). The computer 10 determines the connection destination generation AI based on, for example, whether the wording of the acquired prompt matches or is similar to a group of keywords corresponding to confidential information. Furthermore, the computer 10 inputs the acquired prompt to a classification generation AI that classifies the prompt based on whether or not it contains confidential information, and determines the connection destination generation AI based on the response. Furthermore, the computer 10 inputs the acquired prompt to a determination generation AI that determines an appropriate connection destination generation AI, and determines the connection destination generation AI based on the response.

[0025] A specific example of when the computer 10 determines the connection destination generation AI will be described. The computer 10 performs various language analysis processes for the acquired prompt, such as summarizing confidential information and generating keywords. The computer 10 vectorizes the summaries and keywords of confidential information, compares the vectorized data with vector data of confidential information registered in advance, and performs processes such as identifying summaries of registered file entities corresponding to the acquired prompt, thereby analyzing and identifying the contents of confidential information included in the acquired prompt. The computer 10 generates a prompt using the results of this analysis and identification, and inputs this generated prompt to the classification generation AI and the judgment generation AI. The computer 10 determines the connection destination generation AI based on the answers from the classification generation AI and the judgment generation AI.

[0026] The computer 10 inputs a prompt to the connection destination generation AI based on the determination result (step S3). If the computer 10 determines that the destination generation AI is an on-premise generation AI, it outputs a prompt to the on-premise computer 4. The on-premise computer 4 receives this prompt and inputs the received prompt to the on-premise generation AI. On the other hand, if the computer 10 determines that the connection destination generation AI is a cloud-based generation AI, it outputs a prompt to the cloud computer 5. The cloud computer 5 receives this prompt and inputs the received prompt to the cloud-based generation AI.

[0027] The computer 10 acquires an answer to the prompt from the connection destination generation AI that input the prompt (step S4). The on-premise computer 4 and cloud computer 5 to which the prompt is input generate an answer to the prompt using the functions of the on-premise generation AI and cloud generation AI, and output the generated answer to the computer 10. The computer 10 receives this response and acquires a response to the prompt from the connection destination generation AI that input the prompt.

[0028] The computer 10 outputs the answer (step S5). The computer 10 outputs the acquired answer to the user terminal 3. The user terminal 3 receives this response and displays it on its own display unit or the like. The computer 10 outputs the answer by displaying the answer on the user terminal 3.

[0029] The above is an overview of the information processing system 1. According to the information processing system 1, a prompt is input to an appropriate generation AI, either an on-premise generation AI or a cloud-based generation AI, and an answer is obtained. For example, if the prompt contains confidential information, the prompt is input to the on-premise generation AI, and if the prompt does not contain confidential information, the prompt is input to the cloud-based generation AI and an answer is obtained. As a result, safety can be maintained without requiring enormous resources, making it possible to achieve both cost and safety.

[0030] [Device configuration] 2 is a block diagram showing the configuration of the information processing system 1. The device configuration of the information processing system 1 will be described with reference to FIG. The information processing system 1 is configured with at least a computer 10 that utilizes a generation AI that generates answers to prompts. In this embodiment, the information processing system 1 is configured with a user terminal 3, an on-premise computer 4, and a cloud computer 5 in addition to the computer 10.

[0031] The information processing system 1 is a system in which a computer 10 is connected to a user terminal 3, an on-premise computer 4, and a cloud computer 5 so as to be able to communicate data with each other via a network 8 such as the Internet, an in-house LAN, Wi-Fi, or VPN. Here, the on-premise computer 4 is connected to the computer 10 so as to be able to communicate data via a closed network environment such as an in-house LAN or VPN. The cloud computer 5 is connected to the computer 10 so as to be able to communicate data via an open network environment such as the Internet or Wi-Fi. The user terminal 3 may be connected to the computer 10 so as to be able to communicate data via a closed network environment, or may be connected to the computer 10 so as to be able to communicate data via an open network environment. In addition, the information processing system 1 may include other terminals and devices in addition to the user terminal 3, on-premise computer 4, cloud computer 5 and computer 10, and the number, types and functions of the other terminals and devices can be designed as appropriate.

[0032] The user terminal 3 is the above-mentioned terminal device used by the user. The user terminal 3 has a CPU (Central Processing Unit), GPU (Graphics Processing Unit), RAM (Random Access Memory), ROM (Read Only Memory), etc. as a terminal control unit, and has devices, etc. as a communication unit to enable communication with other terminals, devices, etc. The user terminal 3 also includes, as an input / output unit, various devices for receiving predetermined inputs and inputting / outputting various data.

[0033] The on-premise computer 4 is implemented as a single computer having a server function and connected to the computer 10 so as to be capable of data communication in a closed network environment (such as an internal LAN or VPN). The on-premise computer 4 includes a CPU, GPU, RAM, ROM and the like as a control unit, and includes a device or the like for enabling communication with other terminals, devices and the like as a communication unit. The on-premise computer 4 has a data storage unit such as a hard disk, semiconductor memory, recording medium, memory card, etc. as a memory unit, and the on-premise generation AI is registered therein. The on-premise computer 4 has various devices as a processing unit that executes various processes, and also has functions such as outputting answers to prompts from the on-premise generation AI.

[0034] Cloud computer 5 is realized by one or more computers that have a server function and are connected to computer 10 so as to be able to communicate data with it in an open network environment (Internet, Wi-Fi, etc.). The cloud computer 5 includes a CPU, GPU, RAM, ROM, etc. as a control unit, and includes a device, etc. that enables communication with other terminals, devices, etc. as a communication unit. The cloud computer 5 includes a data storage unit such as a hard disk, semiconductor memory, recording medium, or memory card as a storage unit, and the cloud-based generation AI is registered therein. The cloud computer 5 has, as a processing unit, various devices that execute various processes, and also has functions such as outputting answers to prompts from the cloud-based generation AI.

[0035] The computer 10 has a server function and may be realized, for example, by a single computer, or may be realized by multiple computers like a cloud computer. The computer 10 has a control unit including a CPU, GPU, RAM, ROM, etc., and a communication unit including a device for enabling communication with other terminals and devices, a prompt acquisition unit for acquiring prompts, a prompt input unit for inputting prompts to a connection destination generation AI based on the judgment result, an answer acquisition unit for acquiring answers to the prompts from the connection destination generation AI that input the prompts, and an answer output unit for outputting answers. The computer 10 includes a data storage unit such as a hard disk, semiconductor memory, recording medium, or memory card as a memory unit. The computer 10 includes, as processing units, various devices that execute various processes, a connection destination determination unit that determines whether the on-premise generation AI or the cloud-based generation AI should be the connection destination generation AI, and the like.

[0036] In the computer 10, the control unit loads a predetermined program, thereby realizing a prompt acquisition module, a prompt input module, an answer acquisition module, and an answer output module in cooperation with the communication unit. In addition, in the computer 10, the control unit reads a predetermined program, and in cooperation with the processing unit, realizes a connection destination determination module, a prompt generation module, a prompt input module, a classification result acquisition module, a classification module, and a determination result acquisition module.

[0037] Hereinafter, each process executed by the information processing system 1 will be described together with the process executed by each of the modules described above. In this specification, each module may execute its processing content as its own function, or may execute its processing content via a predetermined application.

[0038] [Answering process executed by computer 10] The answering process executed by the computer 10 will be described with reference to FIG. The figure is a flowchart of the answer processing executed by computer 10. The answer processing includes details of a prompt acquisition process (step S1) that acquires a prompt, a connection destination determination process (step S2) that determines whether the connection destination generation AI should be an on-premise generation AI or a cloud-based generation AI, a prompt input process (step S3) that inputs the prompt to the connection destination generation AI based on the determination result, an answer acquisition process (step S4) that acquires an answer to the prompt from the connection destination generation AI that input the prompt, and an answer output process (step S5) that outputs the answer.

[0039] The prompt acquisition module acquires a prompt (step S10). Prompts are commands that users can use with generative AI, and can include questions, explanations, instructions, requests, summaries, etc. The type of prompt is not particularly limited, and may be any type input by the user as appropriate. The prompt acquisition module acquires a prompt from the user terminal 3 (see FIG. 4). The prompt acquisition module acquires a prompt from the user via a predetermined UI (User Interface) such as a chatbot that exchanges messages with the user in an interactive format. The user terminal 3 receives a prompt input from the user via a predetermined UI or the like, and transmits the received prompt to the computer 10. The prompt acquisition module receives the prompt and acquires the prompt.

[0040] The prompts acquired by the prompt acquisition module will be described with reference to Fig. 4. The figure is a diagram schematically showing the prompts acquired by the prompt acquisition module. In the figure, a prompt 20 is shown.

[0041] Prompt 20 is an example of a question that reads, "You are a sales department employee. Please summarize your operating profit for this fiscal year and give your forecast for operating profit for next fiscal year." The prompt acquisition module uses this prompt 20 as the original prompt (the prompt entered by the user) and executes the connection destination determination process described later.

[0042] Returning to FIG. 3, the rest of the response processing will be explained. The computer 10 executes a connection destination determination process for determining whether the on-premise generation AI or the cloud-based generation AI is to be the connection destination generation AI (step S11). In the process of step S11, the computer 10 executes a first connection destination determination process and / or a second connection destination determination process, which will be described later, to determine a connection destination generation AI to which a prompt is to be input.

[0043] <First connection destination determination process> The first connection destination determination process executed by the computer 10 will be described with reference to Fig. 5. This figure is a flowchart of the first connection destination determination process.

[0044] The prompt generation module generates a classification prompt (step S100). The classification prompt is a prompt to be input to the classification generation AI of computer 10, and is a prompt for classifying the prompt (original prompt) acquired by the processing of step S10. The classification prompt is a prompt for causing the classification generation AI to respond, for example, as to whether or not confidential information is included in the original prompt, and is composed of the original prompt and a command statement that instructs the classification generation AI to classify this original prompt (see FIG. 6).

[0045] Another specific example of the execution of the prompt generation module will now be described. The prompt generation module performs various language analysis processes on the received prompt, such as summarizing the confidential information and generating keywords. The prompt generation module vectorizes the processing results and compares the vectorized processing results with vector data of the confidential information registered in advance. Based on the comparison results, the prompt generation module performs the identification of a summary of the registered file entity corresponding to the obtained prompt, and performs the analysis and identification of the contents of the confidential information included in the obtained prompt. The prompt generation module generates classification prompts based on the analysis and identification results.

[0046] The classification generation AI is a function of the computer 10, and is a generation AI that classifies whether or not an original prompt contains confidential information as an answer to a classification prompt. The specific process by which the classification generation AI classifies the original prompt may be the same as that of a general generation AI, which outputs an answer to an input prompt, or may be a process in which, based on the similarity between the original prompt included in the classification prompt and a preset group of keywords corresponding to confidential information, the original prompt is classified as containing confidential information if it matches or is similar, and the original prompt is classified as not containing confidential information if it does not match or is similar, and each of these is output as an answer.

[0047] The classification prompt is not limited to the above example, but can be designed as appropriate. Furthermore, the classification generation AI is not limited to the above-mentioned example, and can be designed as appropriate. Furthermore, the classification generation AI may be configured so that the computer 10 has the function thereof, or the on-premise computer 4 has the function thereof.

[0048] The classification prompt generated by the prompt generation module will be described with reference to Fig. 6. This figure is a diagram schematically showing the classification prompt generated by the prompt generation module. In this figure, a classification prompt 30 is shown.

[0049] The classification prompt 30 is composed of an original prompt 31 and a command statement 32 . Original prompt 31 is a prompt entered by the user, and is composed of a prompt indicating what the user entered and the prompt that was actually entered. Original prompt 31 is, "The prompt entered by the user is, 'You are an employee of the sales department. Please provide a summary of this fiscal year's operating profit and your forecast for next fiscal year's operating profit.'" Of these, the prompt indicating what the user entered corresponds to the phrase, "The prompt entered by the user is...". The prompt actually entered corresponds to the phrase sandwiched between the prompts indicating what the user entered ('You are an employee of the sales department. Please provide a summary of this fiscal year's operating profit and your forecast for next fiscal year's operating profit.'). The command statement 32 is a prompt for causing the classification generation AI to classify the original prompt 31, and is composed of a command statement for classifying the original prompt 31. An example of the command statement is "Please classify this prompt based on whether it contains confidential information and provide an answer."

[0050] Returning to FIG. 5, the first connection destination determination process will be continued. The prompt input module inputs a classification prompt to the classification generation AI (step S101), and the classification result acquisition module acquires the classification result response (step S102). The prompt input module inputs the generated classification prompt to the classification generation AI. The classification generation AI classifies the original prompt based on the classification prompt as to whether it contains confidential information or not, and outputs the classification result as an answer. The classification result acquisition module acquires an answer and obtains the answer of the classification result of the original prompt.

[0051] The classification module classifies the original prompt based on the obtained classification result (step S103). The classification module classifies whether the original prompt contains confidential information based on the answer output by the classification generation AI.

[0052] For example, if operating profit is confidential information, the answer output by the classification generation AI is that the original prompt contains confidential information, and therefore the classification module classifies the original prompt 31 as a prompt containing confidential information. On the other hand, if operating profit is not confidential information, the answer output by the classification generation AI is that the original prompt does not contain confidential information, so the classification module classifies the original prompt 31 as a prompt that does not contain confidential information.

[0053] The connection destination determination module determines the connection destination generation AI based on the classification result (step S104). If the original prompt is classified as containing confidential information, the connection destination determination module determines that the connection destination generation AI is an on-premise generation AI, and if the original prompt is classified as not containing confidential information, the connection destination determination module determines that the connection destination generation AI is a cloud-based generation AI.

[0054] The computer 10 uses the determined connection destination generation AI to execute the processes from step S12 onwards.

[0055] The above is the first connection destination determination process.

[0056] <Second connection destination determination process> The second connection destination determination process executed by the computer 10 will be described with reference to Fig. 7. This figure is a flowchart of the second connection destination determination process.

[0057] The prompt generation module generates a decision prompt (step S200). The determination prompt is a prompt to be input to the determination generation AI possessed by computer 10, and is a prompt for determining an appropriate connection destination generation AI as an input destination for the prompt (original prompt) acquired by the processing of step S10. The determination prompt is a prompt for determining, for example, whether an on-premise generation AI or a cloud-based generation AI is more appropriate as an appropriate connection destination generation AI to be the input destination, and is composed of the original prompt and a command statement that instructs the determination generation AI to determine an appropriate connection destination generation AI as an input destination for this original prompt (see Figure 8).

[0058] Another specific example of the execution of the prompt generation module will now be described. The prompt generation module performs various language analysis processes on the received prompt, such as summarizing the confidential information and generating keywords. The prompt generation module vectorizes the processing results and compares the vectorized processing results with vector data of the confidential information registered in advance. Based on the comparison results, the prompt generation module performs the identification of a summary of the registered file entity corresponding to the obtained prompt, and performs the analysis and identification of the contents of the confidential information included in the obtained prompt. The prompt generation module generates a prompt for judgment based on the analysis and identification results.

[0059] The judgment generation AI is a function of computer 10, and is a generation AI that determines a connection destination generation AI that can generate an appropriate answer as a response to a judgment prompt. The specific process by which the judgment generation AI determines an appropriate connection destination generation AI as an input destination of the original prompt may be a process in which a general generation AI outputs an answer to the input prompt, or may be a process in which, based on the similarity between the original prompt included in the judgment prompt and a group of keywords corresponding to predetermined confidential information, if there is a match or approximation, the on-premise generation AI is determined to be an appropriate connection destination generation AI, and if there is a mismatch or approximation, the cloud-based generation AI is determined to be an appropriate connection destination generation AI, and output each of these as an answer.

[0060] The prompt for determination is not limited to the above example, but can be designed as appropriate. Furthermore, the judgment generation AI is not limited to the above-mentioned example, and can be designed as appropriate. Furthermore, the judgment generation AI may be configured such that the computer 10 has the function thereof, or the on-premise computer 4 has the function thereof.

[0061] The decision prompt generated by the prompt generation module will be described with reference to Fig. 8. The figure is a diagram schematically showing the decision prompt generated by the prompt generation module. In the figure, a decision prompt 40 is shown.

[0062] The decision prompt 40 is composed of an original prompt 41 and a command statement 42 . The original prompt 41 is a prompt entered by the user, and is composed of a prompt indicating what the user entered and the prompt that was actually entered. The original prompt 41 is, "The prompt entered by the user is, 'You are an employee of the sales department. Please provide a summary of the operating profit for this fiscal year and your forecast for operating profit for next fiscal year.'" Of these, the prompt indicating what the user entered corresponds to the phrase, "The prompt entered by the user is...". The prompt actually entered corresponds to the phrase sandwiched between the prompts indicating what the user entered ('You are an employee of the sales department. Please provide a summary of the operating profit for this fiscal year and your forecast for operating profit for next fiscal year.'). The command statement 42 is a prompt for the judgment generation AI to determine an appropriate destination generation AI as an input destination for the original prompt 41, and is composed of a command statement for determining an appropriate destination generation AI as an input destination for the original prompt 41. The command statement is, "For this prompt, please answer which of the generation AI on OpenAI or the generation AI on the in-house server will provide a higher answer accuracy and density."

[0063] Returning to FIG. 7, the second connection destination determination process will be continued. The prompt input module inputs a prompt for judgment to the judgment generation AI (step S201), and the judgment result acquisition module acquires the judgment result (step S202). The prompt input module inputs the generated decision prompt to the decision generation AI. The judgment generation AI determines an appropriate connection destination generation AI based on the judgment prompt and whether the original prompt contains confidential information, and outputs the determination result as an answer. If the original prompt contains confidential information, the judgment generation AI determines that the appropriate connection destination generation AI is an on-premise generation AI, and if the original prompt does not contain confidential information, the judgment generation AI determines that the appropriate connection destination generation AI is a cloud-based generation AI.

[0064] For example, if operating profit is confidential information, the judgment generation AI determines that the connection destination generation AI that is appropriate as an input destination for the original prompt 41 is an on-premise generation AI. On the other hand, if the operating profit is not confidential information, the judgment generation AI determines that the connection destination generation AI that is appropriate as an input destination for the original prompt 41 is the cloud-based generation AI. The judgment generation AI outputs the judgment result as the answer. The judgment result acquisition module acquires the answer and obtains the judgment result of the original prompt.

[0065] The connection destination determination module determines the connection destination generation AI based on the determination result (step S203). If the determination result of the original prompt is an on-premise generation AI, the connection destination determination module determines that the connection destination generation AI is an on-premise generation AI, and if the determination result of the original prompt is a cloud-based generation AI, the connection destination determination module determines that the connection destination generation AI is a cloud-based generation AI.

[0066] The computer 10 uses the determined connection destination generation AI to execute the processes from step S12 onwards.

[0067] The above is the second connection destination determination process.

[0068] The computer 10 may execute either the first connection destination determination process or the second connection destination determination process, or may execute both processes simultaneously. When both are executed simultaneously, the computer 10 may determine the connection destination generation AI based on both the classification result and the judgment result.

[0069] When the computer 10 determines that the classification result and the judgment result are the same for a connection destination generation AI, it may determine this connection destination generation AI as the input destination of the original prompt.

[0070] On the other hand, if the classification result and the judgment result of the connection destination generation AI are different, the on-premise generation AI may be prioritized as the connection destination generation AI. Furthermore, in this case, the configuration may be such that the user is asked which connection destination generation AI to use.

[0071] In addition, the computer 10 can also be configured to determine the connection destination generation AI in a manner other than executing the determination of the connection destination generation AI using a generation AI, such as the first connection destination determination process or the second connection destination determination process. This case will be explained.

[0072] First, there is a case where it is determined whether or not confidential information is included in the prompt (original prompt) acquired by the processing in step S10. In this case, the connection destination determination module determines that confidential information is included if there is a match or approximation based on the similarity between the original prompt and a group of keywords corresponding to pre-set confidential information, and determines that the connection destination generation AI is an on-premise generation AI; if there is a mismatch or approximation, it determines that confidential information is not included, and determines that the connection destination generation AI is a cloud-based generation AI.

[0073] In addition, if the user has included in the prompt a statement specifying the destination generation AI, such as whether to use an on-premise generation AI, whether to not use an on-premise generation AI, whether to use a cloud-based generation AI, or whether to not use a cloud-based generation AI, the destination determination module can also be configured to determine the destination generation AI based on this statement. However, in this case, even if the user specifies the destination generation AI, it may be inappropriate (for example, if the AI ​​contains confidential information but also contains wording specifying the use of a cloud-based generation AI), so in order to exclude such cases, it is possible to configure the first destination determination process and the second destination determination process to be performed in parallel.

[0074] Even in such a case, it goes without saying that a configuration in which the first connection destination determination process and / or the second connection destination determination process are combined is also possible.

[0075] Returning to FIG. 3, the rest of the response processing will be explained. The prompt input module inputs a prompt to the connection destination generation AI (step S12), and the answer acquisition module acquires an answer from the connection destination generation AI (step S13). The prompt input module inputs the acquired prompt (original prompt) to the connection destination generation AI determined by the connection destination determination process. The prompt input module sends a prompt to a computer having the function of the destination generation AI (on-premise computer 4 if the destination generation AI is an on-premise generation AI, or cloud computer 5 if the destination generation AI is a cloud-based generation AI) and inputs the prompt to the destination generation AI. The computer having the function of the connection destination generation AI receives the prompt, and the connection destination generation AI generates an answer to this prompt. The computer having the function of the connection destination generation AI sends the generated answer to computer 10. The answer acquisition module receives this answer and acquires the answer from the connection destination generation AI.

[0076] The answer output module outputs the answer (step S14). The answer output module outputs the acquired answer via a predetermined UI such as a chatbot that exchanges messages with the user in an interactive format. The answer output module transmits the answer to the user terminal 3. The user terminal 3 receives this response and displays it on its own display unit via a predetermined UI or the like. The answer output module outputs the answer by displaying the answer on the user terminal 3.

[0077] This completes the response process.

[0078] The above-described means and functions are realized by a computer (including a CPU, an information processing device, and various terminals) reading and executing a predetermined program. The program may be provided, for example, from a computer via a network (Software as a Service (SaaS)) or as a cloud service. The program may also be provided in a form recorded on a computer-readable recording medium. In this case, the computer reads the program from the recording medium, transfers it to an internal or external recording device, records it, and executes it. The program may also be pre-recorded on a recording device (recording medium) and provided to the computer from the recording device via a communication line.

[0079] Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments. Furthermore, the effects described in the embodiments of the present invention are merely a list of the most preferable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the embodiments of the present invention.

[0080] A first aspect disclosed in this embodiment is an information processing system that utilizes a generation AI to generate an answer to a prompt, a prompt acquisition unit that acquires a prompt; A connection destination determination unit that determines whether the prompt should be used as the connection destination generation AI, an on-premise generation AI, or a cloud-based generation AI; a prompt input unit that inputs the prompt to the connection destination generation AI based on the determination result; an answer acquisition unit that acquires an answer to the prompt from the connection destination generation AI that input the prompt; a response output unit that outputs the response; An information processing system is provided.

[0081] A second aspect disclosed in this embodiment includes a classification unit that classifies the prompt; Further provided with The connection destination determination unit determines a connection destination generation AI based on the classification result. The present invention provides an information processing system according to a first aspect.

[0082] A third aspect disclosed in this embodiment includes a classification input unit that inputs the prompt to a classification generation AI that classifies the prompt; a classification result acquisition unit that acquires a response to the classification result of the prompt from the classification generation AI that inputs the prompt; Further provided with the classification unit classifies the prompt based on the acquired answer to the classification result; According to a second aspect, there is provided an information processing system.

[0083] In a fourth aspect disclosed in this embodiment, the connection destination determination unit determines, based on the classification result, if the prompt contains confidential information, the on-premise generation AI as the connection destination generation AI, and if the prompt does not contain confidential information, determines the cloud-based generation AI as the connection destination generation AI. According to a second aspect, there is provided an information processing system.

[0084] In a fifth aspect disclosed in this embodiment, the connection destination determination unit determines a generation AI that generates an appropriate response to the prompt as a connection destination generation AI. The present invention provides an information processing system according to a first aspect.

[0085] A sixth aspect disclosed in this embodiment includes a judgment input unit that inputs the prompt to a judgment generation AI that determines a generation AI that generates an appropriate answer; a judgment result acquisition unit that acquires, from the judgment generation AI that has received the prompt, an answer to the judgment result of the generation AI that generates an appropriate answer; Further provided with The connection destination determination unit determines the connection destination generation AI based on the response of the determination result. An information processing system according to a fifth aspect is provided.

[0086] In a seventh aspect disclosed in this embodiment, the connection destination determination unit determines, based on the determination result, if the prompt contains confidential information, the on-premise generation AI as the connection destination generation AI, and if the prompt does not contain confidential information, the cloud-based generation AI as the connection destination generation AI. An information processing system according to a fifth aspect is provided. [Explanation of symbols]

[0087] 1. Information Processing Systems 3. User terminal 4 On-premise computers 5. Cloud Computer 8 Network 10. Computers 20 prompts 30 Sorting Prompts 31 Original Prompt 32 Imperative sentence 40 Judgment Prompts 41 original prompts 42 Imperative sentence

Claims

1. An information processing system utilizing generative AI to generate answers to prompts, a prompt acquisition unit that acquires a first prompt; a prompt generation unit that generates a second prompt for determining whether the first prompt includes confidential information or not, and whether the first prompt includes confidential information, and whether the second prompt is to be a destination AI, i.e., an on-premise generation AI or a cloud generation AI, based on whether the first prompt includes confidential information or not; a connection destination determination unit that determines whether the first prompt should select an on-premise generation AI or a cloud generation AI as the connection destination generation AI based on an answer to the second prompt; a prompt input unit that inputs the first prompt to a connection destination generation AI based on a determination result; an answer acquisition unit that acquires an answer to the first prompt from the connection destination generation AI that has input the first prompt; a response output unit that outputs the response; An information processing system comprising:

2. a classifier for classifying the first prompt; Further provided with The connection destination determination unit determines a connection destination generation AI based on the classification result. The information processing system according to claim 1 .

3. a classification input unit that inputs the second prompt to a classification generation AI that classifies the first prompt; a classification result acquisition unit that acquires an answer to the classification result of the first prompt from the classification generation AI that has received the second prompt; Further provided with The classification unit classifies the first prompt based on the acquired answer to the classification result. The information processing system according to claim 2 .

4. The connection destination determination unit determines, based on the classification result, if the first prompt includes confidential information, the on-premise generated AI as a connection destination generated AI, and if the first prompt does not include confidential information, determines the cloud generated AI as a connection destination generated AI. The information processing system according to claim 2 .

5. The connection destination determination unit determines that a generation AI that generates an answer to the first prompt is a connection destination generation AI. The information processing system according to claim 1 .

6. a judgment input unit that inputs the second prompt to a judgment generation AI that determines a generation AI that generates an answer; a judgment result acquisition unit that acquires an answer to the judgment result of the generation AI that generates an answer from the judgment generation AI that has input the second prompt; Further provided with The connection destination determination unit determines the connection destination generation AI based on the response of the determination result. The information processing system according to claim 5 .

7. The connection destination determination unit determines, based on the determination result, if the first prompt includes confidential information, the on-premise generated AI as the connection destination generated AI, and if the first prompt does not include confidential information, determines the cloud-based generated AI as the connection destination generated AI. The information processing system according to claim 5 .

8. 1. A computer utilizing generative AI to generate answers to prompts, comprising: a prompt acquisition unit that acquires a first prompt; a prompt generation unit that generates a second prompt for determining whether the first prompt includes confidential information or not, and whether the first prompt includes confidential information, and whether the second prompt is to be a destination AI, i.e., an on-premise generation AI or a cloud generation AI, based on whether the first prompt includes confidential information or not; a connection destination determination unit that determines whether the first prompt should select an on-premise generation AI or a cloud generation AI as the connection destination generation AI based on an answer to the second prompt; a prompt input unit that inputs the first prompt to a connection destination generation AI based on a determination result; an answer acquisition unit that acquires an answer to the first prompt from the connection destination generation AI that has input the first prompt; a response output unit that outputs the response; A computer comprising:

9. 1. A computer-implemented information processing method utilizing generative AI for generating answers to prompts, comprising: obtaining a first prompt; generating a second prompt for determining whether the first prompt should be an on-premise generated AI or a cloud generated AI based on whether the first prompt contains confidential information; determining whether the first prompt should select an on-premise generated AI or a cloud generated AI as the destination generated AI based on an answer to the second prompt; inputting the first prompt into a connection destination generation AI based on the determination result; acquiring an answer to the first prompt from the connection destination generation AI that input the first prompt; outputting the answer; An information processing method comprising:

10. A computer that utilizes generative AI to generate answers to prompts. obtaining a first prompt; generating a second prompt for determining whether the first prompt should be an on-premise generated AI or a cloud generated AI based on whether the first prompt contains confidential information; determining whether the first prompt should select an on-premise generated AI or a cloud generated AI as the destination generated AI based on an answer to the second prompt; inputting the first prompt into a connection destination generation AI based on the determination result; acquiring an answer to the first prompt from the connection destination generation AI that input the first prompt; outputting the answer; A computer-readable program for executing the above.

Images