Discussion support device, discussion support method, and program

Abstract

This system provides users with an easy-to-understand explanation of the intermediate products generated during discussions utilizing multiple generative AIs. [Solution] In a discussion support device that is communicatively connected to multiple generating AIs, a setting information acquisition means acquires setting information set by the user regarding the discussion. A control means activates the multiple generating AIs based on the setting information and makes them execute the discussion. A history recording means records the content of each generating AI's statements. A contribution information creation means analyzes the statements of each generating AI and creates contribution information indicating their contribution to the outcome of the discussion. An output means outputs the contribution information.

Description

[Technical Field]

[0001] The present disclosure relates to technology utilizing generative AI (Artificial Intelligence). [Background Art]

[0002] In recent years, the introduction of systems that utilize a plurality of generative AIs to generate results requested by users has been progressing. With the development of such systems, an environment that can flexibly respond to the diverse needs of users is being established. Patent Document 1 describes a system in which an answering device having a plurality of large language models (LLMs) appropriately responds to a user's request. [Prior Art Documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent No. 7588752 [Summary of the Invention] [Problem to be Solved by the Invention]

[0004] When utilizing a plurality of generative AIs, a large amount of intermediate products such as character strings and images are produced before a final result is generated. Therefore, when a user wants to know how the result was obtained, which generative AI played a key role, and the like when checking the result, the user needs to interpret a large amount of intermediate products. For example, when intermediate products are in a chat format, the user has to scroll upward from the finally displayed result to find portions related to the result and generative AIs that contributed, which causes a problem of taking time and effort.

[0005] One object of the present disclosure is to provide a user with the content of intermediate products generated in a discussion utilizing a plurality of generative AIs in an easily understandable manner. [Means for Solving the Problem]

[0006] To address the above issues, from one perspective of this disclosure, the discussion support device is: Connected to communicate with multiple generative AIs, A means for obtaining configuration information to obtain user-defined settings regarding discussions, A control means that activates multiple generating AIs based on the aforementioned configuration information and causes them to perform a discussion, A history recording means for recording the content of each generated AI's statements, A means for creating contribution information that analyzes the statements of each generated AI and creates contribution information that indicates their contribution to the results of the discussion, The system includes an output means for outputting the aforementioned contribution information.

[0007] In other aspects of this disclosure, a discussion support method performed by a discussion support device that is communicatively connected to multiple generative AIs is: Retrieve user-defined settings for the discussion. Based on the aforementioned configuration information, multiple generating AIs are launched and made to perform discussions. Record the content of each generated AI's statements, We analyze the statements of each generating AI and create contribution information that shows their contribution to the results of the discussion. Output the aforementioned contribution information.

[0008] In yet another aspect of this disclosure, a program executed by a discussion support device equipped with a computer and responsively connected to multiple generative AIs is: Retrieve user-defined settings for the discussion. Based on the aforementioned configuration information, multiple generating AIs are launched and made to perform discussions. Record the content of each generated AI's statements, We analyze the statements of each generating AI and create contribution information that shows their contribution to the results of the discussion. The computer is instructed to perform the process of outputting the aforementioned contribution information. [Effects of the Invention]

[0009] According to the present disclosure, the contents of intermediate products generated in a discussion utilizing a plurality of generative AIs can be provided to a user in an easily understandable manner. [Brief Description of the Drawings]

[0010] [Figure 1] An example of a schematic configuration of a discussion support system according to the present disclosure is shown. [Figure 2] It is a block diagram showing an example of the hardware configuration of a server and a user terminal. [Figure 3] It is a block diagram showing an example of the functional configuration of a server. [Figure 4] It is an example of a setting screen. [Figure 5] These are examples of statements by each generative AI. [Figure 6] It is an example of a result screen. [Figure 7] It is an example showing the entire statement history screen [Figure 8] This is an example in which some statements are displayed in a state that can be distinguished from other statements. [Figure 9] It is a flowchart showing an example of screen output processing by a server. [Figure 10] It is a block diagram showing the functional configuration of a discussion support device. [Figure 11] It is a flowchart by a discussion support device. [Mode for Carrying Out the Invention]

[0011] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. [First Embodiment] (Overall Configuration) FIG. 1 is an example of a schematic configuration of a discussion support system 100 to which the discussion support device of the present disclosure is applied. The discussion support system 100 is a system that provides a user with easily understandable visualization of the content of intermediate products generated in discussions utilizing generative AI. Here, the discussion is a multi-person discussion aimed at exchanging opinions and solving problems, and includes debates in the form of clashing opposing opinions.

[0012] In the discussion support system 100 of FIG. 1, a server 1 and a user terminal 2 are communicably connected via a network 5 such as the Internet. A user is a person who receives results of a discussion utilizing generative AI.

[0013] The user terminal 2 is a smartphone, a tablet, a PC, or the like used by a user, and transmits setting information such as a discussion theme, the generative AI to be used, a character to be set, and the like to the server 1, and receives information related to various screens from the server 1.

[0014] The server 1 is an information processing device that performs processing, storage, transmission and reception for various types of data, receives setting information from the user terminal 2, and transmits contribution information indicating intermediate products with high contribution in the discussion to the user terminal 2. The server 1 is connected to a history recording DB 31 which will be described later. The server 1 may be a virtual server existing in a cloud environment. The server 1 is an example of the discussion support device of the present disclosure.

[0015] (Hardware Configuration) FIG. 2(a) is a block diagram showing an example of the hardware configuration of the server 1. As illustrated, the server 1 includes an interface 11, a processor 12, a memory 13, a recording medium 14, a display unit 15, and an input unit 16. These components and the history recording DB 31 are mutually connected via a bus.

[0016] Interface 11 exchanges data with user terminal 2. Interface 11 is used to receive configuration information from user terminal 2 and to send information related to various screens, including contribution information, to user terminal 2.

[0017] Processor 12 is a computer such as a CPU (Central Processing Unit) that controls the entire server 1 by executing pre-prepared programs. Processor 12 can be a CPU, GPU (Graphics Processing Unit), DSP (Digital Signal Processor), MPU (Micro Processing Unit), FPU (Floating Point Number Processing Unit), PPU (Physics Processing Unit), TPU (Tensor Processing Unit), quantum processor, microcontroller, or a combination of these.

[0018] Memory 13 consists of ROM (Read Only Memory), RAM (Random Access Memory), and other components. Memory 13 stores programs executed by the processor 12. Memory 13 is also used as working memory while the processor 12 is executing various processes.

[0019] The recording medium 14 is a non-volatile, non-temporary recording medium such as a disk-shaped recording medium or semiconductor memory, and is configured to be detachable from the server 1. The recording medium 14 stores various programs that the processor 12 executes. When the server 1 performs the screen output processing described later, the programs stored in the recording medium 14 are loaded into the memory 13 and executed by the processor 12.

[0020] The display unit 15, for example, is an LCD (Liquid Crystal Display) and displays various images while the server 1 is operating. The input unit 16 is a keyboard, mouse, touch panel, etc., used by the operator managing the server 1.

[0021] The history log DB31 records the content of each generating AI's statements in a discussion utilizing multiple generating AIs. Specifically, the history log DB31 records history information such as AI identification information to identify the generating AI used in the discussion, the character set for each generating AI, and the statements and times of statements made by each generating AI. The history information may also be recorded in conjunction with the discussion theme and user identification information to identify the users. The information recorded in the history log DB31 can be arbitrarily configured.

[0022] Figure 2(b) is a block diagram showing an example of the hardware configuration of user terminal 2. As shown in the figure, user terminal 2 includes an interface 21, a processor 22, memory 23, recording medium 24, display unit 25, and input unit 26.

[0023] Interface 21 exchanges data with Server 1 via Network 5. Interface 21 is used to send user-configured settings to Server 1 and to receive information related to various screens from Server 1.

[0024] The processor 22 is a computer such as a CPU, and controls the entire user terminal by executing pre-prepared programs. The processor 22 can be a CPU, GPU, DSP, MPU, FPU, PPU, TPU, quantum processor, microcontroller, or a combination of these.

[0025] Memory 23 is composed of ROM, RAM, etc. Memory 23 stores programs executed by the processor 22. Memory 23 is also used as working memory while the processor 22 is executing various processes.

[0026] The recording medium 24 is a non-volatile, non-temporary recording medium such as a disk-shaped recording medium or semiconductor memory, and is configured to be detachable from the user terminal 2. The recording medium 24 stores various programs executed by the processor 22. The display unit 25 is, for example, an LCD, and displays various images while the user operates the user terminal 2. The input unit 26 is, for example, a touch panel, and is used when the user performs a predetermined operation.

[0027] (Functional Configuration) Figure 3 is a block diagram showing an example of the functional configuration of Server 1. Functionally, Server 1 includes a setting information acquisition unit 41, a control unit 42, a history recording unit 43, a contribution information creation unit 44, and an output unit 45.

[0028] The setting information acquisition unit 41, control unit 42, history recording unit 43, contribution information creation unit 44, and output unit 45 are realized by the processor 12 executing a program.

[0029] The configuration information acquisition unit 41 acquires configuration information from the user terminal 2, such as the discussion theme set by the user, the generation AI to be used, and the characters to be set for each generation AI. Figure 4 shows an example of the configuration screen.

[0030] The settings screen, as shown in Figure 4(a), has a theme item and multiple generated AI items and character items. The theme item is where the discussion topic is set. The user can freely enter a theme or select from several predetermined options. As an example, suppose the user sets the theme to "Next-Generation Energy and Social Transformation," as shown in Figure 4(b).

[0031] The Generating AI setting is where users configure the generating AI to be used in discussions. Users can configure the generating AI on the settings screen by selecting from a list of multiple generating AI options or by entering their own desired generating AI. Specifically, the generating AI can be various conversational AIs utilizing LLMs such as ChatGPT, and may even be specially trained in specialized fields such as medicine or law. Note that the generating AI used in discussions is not limited to conversational AIs; any generating AI can be applied, such as a data analysis AI that presents statistical data or numerical evidence.

[0032] The character section is where you set the character for each generated AI. In discussions using multiple generated AIs, assigning different roles to each AI allows for more multifaceted and in-depth discussions. A character is a role assigned to each generated AI; for example, an "engineer" to provide a technical perspective, an "economic analyst" to consider market and economic impacts, or a "doctor" or "lawyer" with specialized knowledge. Users set characters on the settings screen by selecting from a list of characters or by entering a character of their choice. It is also possible to set characters such as a "chairperson" or "moderator" to lead the discussion.

[0033] As shown in Figure 4(a), the generated AI items and character items are paired. For example, suppose the user has configured the system to conduct a discussion using three generated AIs, "AI1" with the character "Engineer," "AI2" with the character "Economic Analyst," and "AI3" with the character "Doctor," as shown in Figure 4(b). Note that while Figure 4(b) uses three generated AIs, it is not limited to this, and the number of generated AIs used can be set arbitrarily.

[0034] The control unit 42 activates multiple generating AIs based on the configuration information and conducts a discussion. Specifically, the control unit 42 initiates the discussion by inputting prompts that include instructions to set a character for each generating AI and instructions to draw a final result by repeatedly exchanging opinions on the theme, based on the configuration information. The instructions requesting the final result can be arbitrarily set depending on the theme.

[0035] Figure 5 shows examples of statements made by each generating AI in a discussion. The history recording unit 43 records the content of statements made by each generating AI in a discussion utilizing multiple generating AIs. Specifically, the history recording unit 43 records the AI ​​identification information that identifies the generating AI used in the discussion, the character set for each generating AI, the statements made by each generating AI, and the time of the statements as history information in the history recording DB 31. The history recording unit 43 may also record the discussion theme and user identification information in association with the history information.

[0036] The contribution information creation unit 44 analyzes the statements of each generating AI based on the history information and creates contribution information that indicates their contribution to the discussion results. The contribution information creation unit 44 includes a result identification unit 51, a starting point identification unit 52, a contribution level calculation unit 53, a summary extraction unit 54, a ranking creation unit 55, and a screen creation unit 56.

[0037] The result identification unit 51 identifies the outcome of the discussion based on the theme included in the setting information and the statements of each generating AI. For example, the result identification unit 51 extracts keywords indicating the outcome, such as "conclusion," "result," and "summary," from the statements of each generating AI, and identifies the result by analyzing the context surrounding the keywords. As another example, the result identification unit 51 creates a prompt requesting the identification of the result based on the theme and the statements of each generating AI, and identifies the result by inputting the prompt into a predetermined generating AI. In the example shown in Figure 5, the result identification unit 51 identifies the result of perspective A as "〇〇 based on △△." Note that the identification of the result is not limited to these examples and can be set arbitrarily.

[0038] The starting point identification unit 52 identifies the starting point statement from the statements of each generating AI that leads to the result. For example, the starting point identification unit 52 identifies statements containing keywords or phrases that directly lead to the result from the statements of each generating AI, and identifies the starting point statement of the result by estimating the causal relationship between preceding and succeeding statements. As another example, the starting point identification unit 52 creates a prompt requesting the identification of the starting point statement of the result from the statements of each generating AI, and identifies the starting point statement by inputting the prompt into a predetermined generating AI. In the example shown in Figure 5, the starting point identification unit 52 identifies statement 70, "I think that...", as the starting point statement. Note that the identification of the starting point statement is not limited to these and can be set arbitrarily. Also, the starting point statement identified by the starting point identification unit 52 is an example of contribution information.

[0039] The contribution calculation unit 53 calculates a contribution score for the result, indicating the degree to which each generated AI's statement contributed. The contribution score is a numerical value, with a larger number indicating a higher contribution and a smaller number indicating a lower contribution. However, it is not limited to this, and the contribution score can be arbitrarily set to be represented by something other than a numerical value such as "high" or "low".

[0040] As an example, the contribution calculation unit 53 calculates the similarity between each generating AI's statement and the result, and calculates the contribution based on the similarity. In this case, the statement with a higher similarity to the result will have a higher calculated contribution. As another example, the contribution calculation unit 53 creates a prompt requesting the calculation of the contribution of each generating AI's statement based on the result, and calculates the contribution of each generating AI's statement by inputting this prompt into a predetermined generating AI. Note that the calculation of contribution is not limited to these methods and can be set arbitrarily. Also, in the example shown in Figure 5, the contribution calculation unit 53 calculates the highest contribution for statement 71, "That's true. But from the perspective of △△, isn't 〇〇 also important?".

[0041] The summary extraction unit 54 extracts a summary of the discussion from each generated AI's statements based on their contribution. As an example, the discussion summary will consist of key statements that contribute to the outcome, as well as the statements immediately before and after them. Specifically, the summary extraction unit 54 extracts the statement with the highest contribution, as well as the statements immediately before and after it, as the summary. In the example shown in Figure 5, the summary extraction unit 54 extracts statement 71 and the statements immediately before and after it as the summary. As another example, the discussion summary may consist of all statements in chronological order, from the starting statement to the statement with the highest contribution. Note that the summary extracted by the summary extraction unit 54 is just one example of contribution information.

[0042] The ranking creation unit 55 creates information indicating the generative AIs that performed well in the discussion. As one example, the ranking creation unit 55 creates a contribution ranking that ranks the generative AIs with the highest contributions. As another example, the ranking creation unit 55 creates information designating the generative AI with the highest contribution and the generative AI that made the initial statement as the Contribution MVP and Initial Statement MVP, respectively.

[0043] Furthermore, the ranking creation unit 55 may create a contribution ranking by ranking characters assigned to high-contributing generation AIs, or it may create information designating the characters assigned to the generation AI with the highest contribution and the generation AI that made the initial statement as Contribution MVP and Initial MVP, respectively. Also, the information created by the ranking creation unit 55 is just one example of contribution information.

[0044] The screen creation unit 56 creates various screens. Figure 6 is an example of a results screen. When the discussion ends, the screen creation unit 56 creates a results screen, as shown in Figure 6(a), which includes the discussion results and a contribution ranking. The results screen is configured so that when the user performs a predetermined operation, for example, hovering the mouse over the underlined part of the discussion results, "Based on △△, ○○", a summary 75 is displayed, as shown in Figure 6(b). Furthermore, the results screen is configured so that when the user performs a predetermined operation, for example, double-clicking the underlined part of the discussion results, "Based on △△, ○○", a comment history screen is displayed, showing the comment history with the highest contribution. Specifically, the string "Based on △△, ○○" in the discussion results is linked to the comment history with the highest contribution.

[0045] Figure 7 is an example showing the entire message history screen. As shown in Figure 7, the message history screen displays all message history, and normally the portion displayed by user terminal 2 is within the dashed line 80. The user can view the entire screen by scrolling up and down using a predetermined operation. When the user double-clicks the underlined part "Based on △△, 〇〇" in the results screen shown in Figure 6(a) of the discussion results, user terminal 2 displays the area within the dashed line 80 of the message history screen shown in Figure 7. The area within the dashed line 80 includes the summary 75 which contains the key points in producing the result, and specifically, as shown in Figure 7, the message with the highest contribution 71 and the summaries 75 before and after it are displayed. Note that the area within the dashed line 80 is not limited to this, and can be arbitrarily set, for example, from the starting message 70 to the messages included in the summary 75.

[0046] As shown in Figure 7, the message history screen displays the starting message 70 and the message with the highest contribution 71 in a way that makes them distinguishable from other messages. Messages included in the summary may also be displayed in a way that makes them distinguishable from other messages. Alternatively, as shown in Figure 8, all messages from the starting message 70 to the messages included in the summary 75 may be displayed in chronological order in a way that makes them distinguishable from other messages.

[0047] When the discussion ends, the output unit 45 sends screen information related to the results screen to the user terminal 2. Furthermore, when the output unit 45 receives a history request from the user on the results screen, such as by double-clicking on a link, it sends screen information related to the conversation history screen to the user terminal 2.

[0048] In the above configuration, the configuration information acquisition unit 41, control unit 42, history recording unit 43, contribution information creation unit 44, and output unit 45 of the server 1 are examples of the configuration information acquisition means, control means, history recording means, contribution information creation means, and output means of the disclosure, respectively.

[0049] (Screen output processing) Next, we will explain the screen output processing by Server 1. Figure 9 is a flowchart showing an example of screen output processing by Server 1. This process is achieved by the processor 12 shown in Figure 2 executing a pre-prepared program.

[0050] First, Server 1 obtains configuration information from User Terminal 2 (Step S101). Next, Server 1 conducts a discussion using multiple Generator AIs based on the configuration information (Step S102). Server 1 records the content of each Generator AI's statements during the discussion as history information in the History Record DB 31 (Step S103).

[0051] Next, Server 1 refers to the history record DB31 and identifies the outcome of the discussion based on the configuration information and the statements of each generating AI (Step S104). Based on the identified outcome, Server 1 identifies the starting statement that led to the result from the statements of each generating AI (Step S105). Server 1 calculates the contribution of each generating AI's statement to the result (Step S106). Based on the contribution, Server 1 extracts a summary of the discussion from the statements of each generating AI (Step S107). Also, Server 1 creates a ranking of the generating AIs in the discussion based on the starting statement and contribution (Step S108).

[0052] Next, Server 1 creates a results screen and sends information about the results screen to User Terminal 2 (Step S109). User Terminal 2 displays the results screen based on the received information. If the user wants to view the conversation history and check the summary of the discussion or the conversation that started the results, they make a history request using a predetermined operation on the results screen. Server 1 determines whether or not it has received a history request from User Terminal 2 (Step S110). If no history request has been received (Step S110; No), Server 1 terminates the screen output process. On the other hand, if a history request has been received (Step S110; Yes), Server 1 creates a conversation history screen and sends information about the conversation history screen to User Terminal 2 (Step S111). Thus, Server 1 terminates the screen output process.

[0053] This discussion support system 100 visualizes and provides to the user the extent to which the content of each generating AI's intermediate output influenced the final discussion result, as contribution information. Furthermore, users can easily access the relevant sections as needed. Specifically, by providing information such as contribution rankings and MVPs, the discussion support system 100 allows users to easily understand which generating AI played a significant role in the discussion. Additionally, by displaying a summary of the discussion along with the results, and clearly showing in the comment history the comments that initiated the results or those with high contributions, the system reduces the time and effort users need to understand the intent behind the results.

[0054] [First variation] In the above embodiment, Server 1 may assign a high contribution to the utterance that serves as the starting point for the result, regardless of its similarity. In this case, the user may set parameters for the utterance that serves as the starting point for the result and for utterances similar to the result in advance through a predetermined operation, and Server 1 may calculate the contribution of each utterance by the generating AI according to the set parameters. This makes it possible to visualize the utterances of the generating AI that played an active role in the discussion according to the parameters set by the user.

[0055] [Second variation] Server 1 may identify statements from each generated AI's statements that have a contribution level above a threshold as relevant statements, and display these relevant statements on the statement history screen in a way that makes them distinguishable from other statements. This makes it possible to easily visualize statements that have a contribution level above a threshold to the outcome of the discussion. Note that the processing of Server 1 in this modified example is just one example of the processing performed by the relevant statement identification means.

[0056] [Second Embodiment] Figure 10 is a block diagram showing an example of the functional configuration of the discussion support device in this disclosure. The discussion support device 90 comprises a setting information acquisition means 91, a control means 92, a history recording means 93, a contribution information creation means 94, and an output means.

[0057] Figure 11 is a flowchart illustrating an example of processing by the discussion support device 90. The discussion support device 90 is connected to multiple generating AIs for communication. The setting information acquisition means 91 acquires setting information set by the user regarding the discussion (step S201). The control means 92 activates multiple generating AIs based on the setting information and executes the discussion (step S202). The history recording means 93 records the content of each generating AI's statements (step S203). The contribution information creation means 94 analyzes the statements of each generating AI and creates contribution information indicating their contribution to the outcome of the discussion (step S204). The output means 95 outputs the contribution information (step S205). The discussion support device 90 makes it possible to visualize the content of intermediate products generated in a discussion utilizing multiple generating AIs in an easy-to-understand manner for the user.

[0058] In addition, some or all of the above embodiments (including modifications, the same applies hereinafter) may also be described as follows, but are not limited to the following.

[0059] (Note 1) Connected to communicate with multiple generative AIs, A means for obtaining configuration information to obtain user-defined settings regarding discussions, A control means that activates multiple generating AIs based on the aforementioned configuration information and causes them to perform a discussion, A history recording means for recording the content of each generated AI's statements, A means for creating contribution information that analyzes the statements of each generated AI and creates contribution information that indicates their contribution to the results of the discussion, Output means for outputting the aforementioned contribution information, A discussion support device equipped with the following features.

[0060] (Note 2) The aforementioned means for creating contribution information is: A result identification means for identifying the results of the aforementioned discussion, A contribution calculation means for calculating a contribution score that indicates the degree to which each generated AI statement contributed to the aforementioned result, The system includes a summary extraction means for extracting a summary of the discussion from the statements of each generating AI based on the contribution level, The output means is the discussion support device described in Appendix 1, which outputs the summary as the contribution information.

[0061] (Note 3) The summary extraction means is a discussion support device as described in Appendix 2, which extracts the statement with the highest contribution and the statements before and after that statement as the summary.

[0062] (Note 4) The contribution calculation means is a discussion support device as described in Appendix 2, which calculates the contribution based on the similarity between the result and the statements made by each generating AI.

[0063] (Note 5) The output means, upon receiving a history request, outputs the content of each generating AI's statements in chronological order. The discussion support device described in Appendix 2 outputs the statements included in the summary in a manner that makes them distinguishable from other statements.

[0064] (Note 6) The contribution information creation means includes a starting point identification means that identifies the starting point statement for the result from the statements of each generating AI, The output means, upon receiving a history request, outputs the content of each generating AI's statements in chronological order. The discussion support device described in Appendix 2 outputs the aforementioned starting statement in a manner that makes it distinguishable from other statements.

[0065] (Note 7) The contribution calculation means is a discussion support device as described in Appendix 2, which calculates the contribution by inputting a prompt to the generating AI, which includes an instruction sentence instructing the generating AI to calculate the contribution of each statement to the result, and the content of each generating AI's statement.

[0066] (Note 8) The contribution calculation means includes a ranking creation means that creates a contribution ranking for each generated AI based on the contribution. The output means is the discussion support device described in Appendix 2, which outputs the contribution ranking as the contribution information.

[0067] (Note 9) A discussion support method performed by a discussion support device that is communicatively connected to multiple generative AIs, Retrieve user-defined settings for the discussion. Based on the aforementioned configuration information, multiple generating AIs are launched and made to perform discussions. Record the content of each generated AI's statements, We analyze the statements of each generating AI and create contribution information that shows their contribution to the results of the discussion. A discussion support method for outputting the aforementioned contribution information.

[0068] (Note 10) A program executed by a discussion support device equipped with a computer and capable of communicating with multiple generative AIs, Retrieve user-defined settings for the discussion. Based on the aforementioned configuration information, multiple generating AIs are launched and made to perform discussions. Record the content of each generated AI's statements, We analyze the statements of each generating AI and create contribution information that shows their contribution to the results of the discussion. A program that causes the computer to perform the process of outputting the aforementioned contribution information.

[0069] (Note 11) The contribution information creation means includes a starting point identification means that identifies the starting statement based on the results, The contribution calculation means is a discussion support device as described in Appendix 4, which calculates the contribution of the starting statement as high regardless of the degree of similarity.

[0070] (Note 12) The discussion support device includes parameter acquisition means for acquiring parameters set for the statement that serves as the starting point for the result and for statements similar to the result. The aforementioned contribution information creation means includes a starting point identification means that identifies the starting statement based on the aforementioned results, The contribution calculation means is a discussion support device as described in Appendix 4, which calculates the contribution of each AI-generated statement based on the parameters.

[0071] (Note 13) The contribution information creation means includes a starting point identification means that identifies the starting statement based on the results, The output means, upon receiving a history request, outputs the content of each generating AI's statements in chronological order, and outputs all statements included in the chronological order from the starting statement to the statements included in the summary, in a state that can be distinguished from other statements, as described in Appendix 2 of the discussion support device.

[0072] (Note 14) The summary extraction means is a discussion support device as described in Appendix 13, which extracts all statements included in a time series, from the starting statement to the statement with the highest contribution, as the summary.

[0073] (Note 15) The system includes a related identification means for identifying statements whose contribution level is above a threshold as related statements, The output means, upon acquiring a history request, outputs the content of each generating AI's statements in chronological order, and the discussion support device described in Appendix 2 outputs the related statements in a manner that makes them distinguishable from other statements.

[0074] Furthermore, some or all of the configurations described in Appendices 2-8 and 11-15, which are dependent on Appendice 1 above, may also be dependent on Appendices 9 and 10 in the same way as Appendices 2-8 and 11-15. Moreover, not limited to Appendices 1, 9, and 10, some or all of the configurations described as appendices may also be dependent on various hardware, software, various recording means for recording software, or systems, without departing from the embodiments described above.

[0075] While the present disclosure has been described above with reference to embodiments, the present disclosure is not limited to the embodiments described above. Various modifications to the structure and details of the present disclosure may be understood by those skilled in the art within the scope of the present disclosure. That is, the present disclosure includes the entire disclosure, including the claims, and of course, various modifications and alterations that those skilled in the art may make in accordance with the technical idea. [Explanation of Symbols]

[0076] 1 server 2 User terminals 11, 21 Interfaces 12, 22 processors 13, 23 memory 14, 24 Recording media 15, 25 Display section 16, 26 Input section 31 History Record Database 41 Configuration Information Acquisition Unit 42 Control Unit 43 History Record Section 44 Contribution Information Creation Department 45 Output section 100 Discussion Support Systems

Claims

1. Connected to communicate with multiple generating AIs, A means for obtaining configuration information to obtain user-defined settings regarding discussions, A control means that activates multiple generating AIs based on the aforementioned configuration information and causes them to perform a discussion, A history recording means for recording the content of each generated AI's statements, A means for creating contribution information that analyzes the statements of each generated AI and creates contribution information that indicates their contribution to the results of the discussion, Output means for outputting the aforementioned contribution information, A discussion support device equipped with the following features.

2. The aforementioned means for creating contribution information is: A result identification means for identifying the results of the aforementioned discussion, A contribution calculation means that calculates a contribution score indicating the degree to which each AI-generated statement contributed to the aforementioned result, The system includes a summary extraction means for extracting a summary of the discussion from the statements of each generated AI based on the contribution level, The discussion support device according to claim 1, wherein the output means outputs the summary as the contribution information.

3. The discussion support device according to claim 2, wherein the summary extraction means extracts the statement with the highest contribution and the statements before and after that statement as the summary.

4. The discussion support device according to claim 2, wherein the contribution calculation means calculates the contribution based on the similarity between the result and the statements made by each generating AI.

5. The output means, upon receiving a history request, outputs the content of each generated AI's statements in chronological order. The discussion support device according to claim 2, which outputs the statements included in the summary in a manner that can be distinguished from other statements.

6. The contribution information creation means includes a starting point identification means that identifies the starting point statement for the result from the statements of each generating AI, The output means, upon receiving a history request, outputs the content of each generated AI's statements in chronological order. The discussion support device according to claim 2, which outputs the starting statement in a manner that can be distinguished from other statements.

7. The discussion support device according to claim 2, wherein the contribution calculation means calculates the contribution by inputting a prompt to the generating AI, which includes an instruction statement instructing the generating AI to calculate the contribution of each statement to the result, and the content of each generating AI's statement.

8. The contribution calculation means includes a ranking creation means that creates a contribution ranking for each generated AI based on the contribution. The discussion support device according to claim 2, wherein the output means outputs the contribution ranking as the contribution information.

9. A discussion support method performed by a discussion support device that is communicatively connected to multiple generating AIs, Retrieve user-defined settings for the discussion. Based on the aforementioned configuration information, multiple generating AIs are activated and a discussion is performed. Record the content of each generated AI's statements, The statements of each generated AI are analyzed, and contribution information is created that shows their contribution to the results of the discussion. A discussion support method for outputting the aforementioned contribution information.

10. A program executed by a discussion support device equipped with a computer and capable of communicating with multiple generating AIs, Retrieve user-defined settings for the discussion. Based on the aforementioned configuration information, multiple generating AIs are activated and a discussion is performed. Record the content of each generated AI's statements, The statements of each generated AI are analyzed, and contribution information is created that shows their contribution to the results of the discussion. A program that causes the computer to perform the process of outputting the aforementioned contribution information.

Images