Creation support apparatus, creation support method, and program

Abstract

To smoothly proceed with a discussion on a predetermined theme using a generative AI.SOLUTION: A discussion support apparatus 4 includes: a persona setting unit 131 that provides a generative AI with two or more persona-setting prompts; an utterance acquisition unit 431 that acquires a prompt using a moderator-prompt template and a theme, provides the prompt to the generative AI, and acquires utterance information from each of one or more agents among two or more agents; an output unit 44 that outputs one or more items of the utterance information; and a moderator determination unit 133 that determines whether moderator utterance conditions are satisfied. When it is determined that the moderator utterance conditions are satisfied, the utterance acquisition unit 431 provides the generative AI with a prompt for each of one or more agents other than the moderator and acquires one or more items of utterance information. By setting a moderator agent by a discussion support apparatus, the progress of a discussion on a predetermined theme using a generative AI becomes smooth.SELECTED DRAWING: Figure 20

Description

[Technical Field]

[0001] The present invention relates to a discussion support device that uses a generation AI to hold discussions on a predetermined topic. [Background technology]

[0002] Conventionally, there has been an idea creation support system that is said to be a platform that provides an environment in which participants can easily create ideas (see Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2023-35936 Summary of the Invention [Problem to be solved by the invention]

[0004] However, with conventional technology, it has been difficult to smoothly advance discussions on a given topic, such as generating ideas, using generative AI. [Means for solving the problem]

[0005] The first invention of the conference support device is a conference support device for two or more parties, including two or more agents of a generation AI, to hold a conference on a predetermined theme, and includes a prompt management unit that stores setting prompt templates for configuring prompts for character settings of two or more agents, including a moderator's agent, and one or more moderator prompt templates, which are templates for configuring prompts to be given to the generation AI for the progress of the conference on the theme and are templates for the moderator; an attribute value receiving unit that receives the theme; a character setting unit that obtains two or more setting prompts using the setting prompt templates and gives each of the one or more setting prompts to the generation AI; and The discussion support device comprises a utterance acquisition unit that acquires a prompt using a moderator prompt template, provides the prompt to the generation AI, and acquires utterance information from one or more of the two or more agents; an output unit that outputs one or more utterance information acquired by the utterance acquisition unit; and a moderator judgment unit that determines whether or not a moderator utterance condition, which is a condition for the moderator's agent to speak and is a condition related to the prompt provided to the generation AI or a condition related to an answer from the generation AI, is met.When the moderator judgment unit determines that the moderator utterance condition is met, the utterance acquisition unit acquires a prompt based on a moderator prompt template, provides the prompt to the generation AI, and acquires one or more utterance information.

[0006] With this configuration, by setting a moderator agent, the progress of discussions on a predetermined topic can be made smooth using the generation AI.

[0007] Furthermore, the second invention of the conference support device differs from the first invention in that the moderator's statement condition is that the statement acquisition unit has acquired statement information from all two or more parties.

[0008] With this configuration, by setting a moderator agent, the progress of discussions on a predetermined topic can be made smooth using the generation AI.

[0009] Furthermore, the consultation support device of the third invention, compared to the first or second invention, further includes an agent determination unit that determines one or more agents according to a theme, and the character setting unit obtains setting prompts for the one or more agents determined by the agent determination unit and provides the setting prompts to the generation AI.

[0010] With this configuration, the generation AI can be used to automatically set appropriate agents to advance discussions on a given topic.

[0011] The creation support device of the present invention is a creation support device that supports idea creation by two or more parties including one or more agents of a generation AI, and is equipped with: a prompt management unit that stores setting prompt templates for constructing prompts for character settings of one or more agents, and phase prompt templates for constructing prompts to be given to the generation AI in each of two or more phases including an idea generation phase and an idea selection phase; an attribute value storage unit that stores one or more idea attribute values ​​that are attribute values ​​related to the idea to be created; a character setting unit that acquires one or more setting prompts using the setting prompt template and gives each of the one or more setting prompts to the generation AI; an idea acquisition unit that acquires prompts using a phase prompt template corresponding to the idea generation phase and one or more idea attribute values, gives the prompts to one or more agents of the generation AI, and acquires two or more ideas; an idea selection unit that acquires prompts using a phase prompt template corresponding to the idea selection phase and two or more ideas acquired by the idea acquisition unit, gives the prompts to the generation AI, and acquires one or more ideas selected based on the two or more ideas acquired by the idea acquisition unit; and an idea output unit that outputs the one or more ideas acquired by the idea selection unit.

[0012] This configuration can assist in creating appropriate ideas using generative AI.

[0013] In addition, the creation support device of the present invention, compared to the above invention, further includes an idea receiving unit that receives one or more ideas from a person, and the idea selection unit obtains a prompt using a phase prompt template corresponding to the idea selection phase, two or more ideas obtained by the idea acquisition unit, and one or more ideas received by the idea reception unit, provides the prompt to the generation AI, and obtains one or more ideas selected based on the two or more ideas obtained by the idea acquisition unit and the one or more ideas received by the idea reception unit.

[0014] This configuration can help people and generative AI collaborate to create appropriate ideas.

[0015] In addition, the creation support device of the present invention, compared to the above invention, further includes an analysis unit that determines whether the ideas received by the idea receiving unit satisfy the consideration conditions and acquires the judgment result, and the idea selection unit acquires a prompt using a phase prompt template corresponding to the idea selection phase and two or more ideas excluding ideas that are determined not to satisfy the consideration conditions, provides the prompt to the generation AI, and acquires one or more ideas selected based on the two or more ideas acquired by the idea acquisition unit and the one or more ideas received by the idea receiving unit.

[0016] This configuration can eliminate input from inappropriate people.

[0017] Furthermore, in the creation support device of the present invention, in addition to the above invention, the creation support device further comprises an error output unit that outputs an error message when the analysis unit determines that the review conditions are not satisfied.

[0018] With this configuration, if an input is made by an inappropriate person, the person can be notified that the input is inappropriate.

[0019] Furthermore, in the creation support device of the present invention, compared to the above invention, the analysis unit analyzes each of one or more ideas received by the idea receiving unit to obtain an error class, and the error output unit outputs an error message including the error class.

[0020] With this configuration, when an input is made by an inappropriate person, the person can be more appropriately notified that the input is inappropriate.

[0021] In addition, the creation support device of the present invention is a creation support device in which, compared to the above invention, there are two or more agents, one of the two or more agents is a moderator's agent, the prompt management unit stores a setting prompt template for the moderator and a phase prompt template for the moderator, the character setting unit also provides the setting prompt for the moderator to the generation AI, the idea acquisition unit corresponds to the idea generation phase, acquires a prompt using the phase prompt template for the moderator and one or more idea attribute values, provides the prompt to the generation AI, and acquires ideas from one or more of the two or more agents, and the idea selection unit corresponds to the idea selection phase, acquires a prompt using the phase prompt template for the moderator and two or more ideas acquired by the idea acquisition unit, provides the prompt to the generation AI, and acquires one or more ideas selected based on the two or more ideas acquired by the idea acquisition unit.

[0022] With this configuration, the moderator agent can help create more appropriate ideas using the generative AI.

[0023] In addition, the creation support device of the present invention, compared to the above invention, further includes a moderator judgment unit that determines whether the moderator's remark conditions, which are conditions for the moderator's agent to speak and are conditions regarding the prompt given to the generation AI or conditions regarding the answer from the generation AI, are met, and when the moderator judgment unit determines that the moderator's remark conditions are met, the idea acquisition unit corresponds to the idea generation phase, acquires prompts for one or more agents other than the moderator, gives the prompts to the generation AI, and acquires two or more ideas.

[0024] With this configuration, the moderator agent can provide appropriate prompts to help generate more appropriate ideas using the generative AI.

[0025] Furthermore, the creation support device of the present invention is a creation support device in which, compared to the above invention, the moderator's comment condition is that the idea acquisition unit has acquired ideas from all two or more parties excluding the moderator.

[0026] With this configuration, the moderator agent can provide appropriate prompts to the generation AI at appropriate times.

[0027] In addition, the creation support device of the present invention is a creation support device that, compared to the above invention, further includes a current management unit that stores a current identifier that identifies the current phase, which is one of two or more phases including an idea generation phase and an idea selection phase, and an output unit that outputs the answer passed from the generation AI in a manner that allows the current identified by the current identifier to be grasped.

[0028] This configuration makes it possible to clearly present the flow of discussion for generating appropriate ideas using generative AI.

[0029] Moreover, in the creation support device of the present invention, in addition to the above invention, an attribute value receiving unit that receives a theme, and one idea attribute value among the one or more idea attribute values ​​is the theme.

[0030] With this configuration, the generative AI can be used to assist in generating ideas related to accepted themes.

[0031] Furthermore, the creation support device of the present invention, in contrast to the above invention, further includes an attribute value receiving unit that receives a theme, and an agent determination unit that determines one or more agents according to the theme, and the character setting unit obtains setting prompts for the one or more agents determined by the agent determination unit and provides the setting prompts to the generation AI.

[0032] With this configuration, it is possible to automatically set appropriate agents to generate appropriate ideas using generative AI.

[0033] In addition, the creation support device of the present invention is a creation support device that, compared to the above invention, further includes a score acquisition unit that performs natural language processing on two or more ideas acquired by the idea acquisition unit or the idea selection unit, and acquires a score related to the diversity of the ideas, and a score output unit that outputs the score.

[0034] This configuration makes it possible to evaluate the diversity of a set of ideas obtained using a generative AI.

[0035] In addition, the creation support device of the present invention is different from the above invention in that each of the two or more phases includes an idea generation phase, an idea criticism phase, and an idea selection phase, and further includes an idea criticism unit that assigns the idea acquired by the idea acquisition unit to a variable of a criticism prompt template, which is a phase prompt template corresponding to the idea criticism phase, constructs a criticism prompt, passes the criticism prompt to the generation AI, and acquires one or more criticisms from the generation AI, and after the idea criticism unit acquires one or more criticisms, the idea selection unit acquires a prompt using the phase prompt template corresponding to the idea selection phase and two or more ideas acquired by the idea acquisition unit, gives the prompt to the generation AI, and acquires one or more ideas selected based on the two or more ideas acquired by the idea acquisition unit.

[0036] This configuration can help create more effective ideas using generative AI.

[0037] In addition, the creation support device of the present invention is a creation support device that, in addition to the above-mentioned invention, further includes a score acquisition unit that acquires a score for each of the two or more ideas acquired by the idea acquisition unit or the idea selection unit using one or more criticisms corresponding to the idea, and the one or more criticisms acquired by the idea criticism unit, and a score output unit that outputs the score for each of the two or more ideas.

[0038] With this configuration, it is possible to evaluate each of two or more ideas obtained using the generation AI.

[0039] In addition, compared to the above invention, the creation support device of the present invention is a creation support device that further includes a transition management unit that stores transition information that identifies the transition of each of two or more phases including an idea generation phase and an idea selection phase, one or more transition conditions that are conditions under which a phase transitions and that correspond to the phase before the transition and the phase after the transition, and a control unit that determines whether the current phase, Current, matches the transition condition corresponding to it, and if the transition condition is matched, changes Current from the current phase to the post-transition phase that corresponds to the transition condition.

[0040] This configuration makes it possible to appropriately control the phases that make up the flow of discussion to generate appropriate ideas using generative AI. [Effects of the Invention]

[0041] According to the discussion support device of the present invention, by setting a moderator agent, discussions on a predetermined topic can proceed smoothly using a generating AI.

[0042] Furthermore, the creation support device according to the present invention can support the creation of appropriate ideas using a generation AI. [Brief explanation of the drawings]

[0043] [Figure 1] Conceptual diagram of information system A in embodiment 1 [Figure 2] Block diagram of Information System A [Figure 3] Block diagram of the creation support device 1 [Figure 4] A flowchart illustrating an example of the operation of the creation support device 1. [Figure 5] A flowchart illustrating an example of the preprocessing [Figure 6] A flowchart illustrating an example of the agent determination process [Figure 7] Flowchart illustrating an example of same person setting processing [Figure 8]Flowchart illustrating an example of viewpoint acquisition processing [Figure 9] A flowchart illustrating an example of the idea acquisition process [Figure 10] A flowchart illustrating an example of the idea analysis process [Figure 11] A flowchart illustrating an example of the idea criticism process. [Figure 12] A flowchart illustrating an example of the idea selection process [Figure 13] FIG. 1 shows the transition management table. [Figure 14] Figure showing the prompt management table [Figure 15] A diagram showing an example of the agent information [Figure 16] Figure showing an example of the output [Figure 17] Figure showing an example of the output [Figure 18] Figure showing an example of the output [Figure 19] Figure showing an example of the output [Figure 20] Block diagram of a conference support device 4 according to a second embodiment. [Figure 21] A flowchart illustrating an example of the operation of the conference support device 4 [Figure 22] A flowchart illustrating an example of the preprocessing [Figure 23] Flowchart illustrating an example of the comment acquisition process [Figure 24] Block diagram of a computer system according to the above embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0044] Hereinafter, embodiments of the creation support device and the like will be described with reference to the drawings. Note that, in the embodiments, components with the same reference numerals perform similar operations, and therefore, repeated description may be omitted.

[0045] (Embodiment 1) In this embodiment, we will describe a creation support device that supports the creation of ideas by two or more parties including one or more generation AI agents, and that acquires two or more ideas in an idea generation phase, and then acquires and outputs one or more ideas from the two or more acquired ideas in a subsequent idea selection phase.

[0046] In this embodiment, a creation support device that allows people to participate in idea creation will be described.

[0047] In this embodiment, a creation support device is described that checks input from a person and, depending on the result of the check, may or may not provide the input to the generation AI. In this embodiment, if the check result of the input from a person is an error, an error message or an error class is output.

[0048] In this embodiment, a creation support device in which a moderator agent exists will be described.

[0049] In this embodiment, a creation support device will be described that operates so that the moderator's agent gives a prompt to the other agents when a predetermined condition is met.

[0050] In addition, in this embodiment, a creation support device will be described that outputs answers, etc. from a generation AI for creating ideas in a manner that allows the current phase to be known within the overall phases of creating ideas.

[0051] In this embodiment, a creation support device that receives a theme from a user and creates ideas according to the theme will be described.

[0052] In this embodiment, a creation support device that automatically determines necessary agents will be described.

[0053] In this embodiment, a creation support device that evaluates a set of output ideas from the viewpoint of diversity will be described.

[0054] Furthermore, in this embodiment, a creation support device will be described that manages two or more phases and phase transition conditions until an idea is created, and controls the phase transition based on the managed information.

[0055] In this specification, information X being associated with information Y means that information Y can be obtained from information X, or information X can be obtained from information Y, and the method of association is not important. Information X and information Y may be linked, may exist in the same buffer, information X may be included in information Y, or information Y may be included in information X, etc.

[0056] Furthermore, in this specification, selecting or determining information Z means obtaining information Z, obtaining a pointer to information Z, obtaining the ID of information Z, setting a flag on information Z, etc., and it is sufficient if information Z can be accessed.

[0057] FIG. 1 is a conceptual diagram of information system A in this embodiment. Information system A comprises a creation support device 1, one or more terminal devices 2, and one or more generating AI devices 3. The program stored in generating AI device 3 is referred to as generating AI. The generating AI device 3 may also be referred to as generating AI. Information system A may have one generating AI device 3, or two or more generating AI devices 3. If there are two or more generating AI devices 3, it is preferable that there be a generating AI device 3 corresponding to each agent described below. The generating AI devices are external devices to information system A, and information system A may be considered not to have a generating AI device 3.

[0058] The creation support device 1 is, for example, a cloud server or an ASP server, but the type does not matter. The creation support device 1 may also be a terminal. If the creation support device 1 is a terminal, the information system A does not require a terminal device 2. If the creation support device 1 is a terminal, it can be considered that the creation support device 1 also serves as the terminal device 2. If the creation support device 1 is a terminal, it is, for example, a so-called personal computer, smartphone, or tablet terminal, but the type does not matter.

[0059] The creation support device 1 may have one or more generation AIs. In such a case, the information system A does not need to have the generation AI device 3.

[0060] The terminal device 2 is a terminal used by a user. The user is a person who uses the creation support device 1. The user may be a person who inputs an idea, or may be a person who checks an idea output by the creation support device 1.

[0061] The generation AI device 3 is, for example, a ChatGPT (registered trademark) server or a Bard (registered trademark) server.

[0062] It is preferable that the creation support device 1 and one or more terminal devices 2, and the creation support device 1 and one or more generation AI devices 3, can communicate with each other via a network such as the Internet.

[0063] 2 is a block diagram of the information system A according to the present embodiment, and FIG.

[0064] The creation support device 1 includes a storage unit 11, a reception unit 12, a processing unit 13, and an output unit 14. The storage unit 11 includes a transition management unit 111, a prompt management unit 112, an attribute value storage unit 113, and a current management unit 114. The reception unit 12 includes an attribute value reception unit 121 and an idea reception unit 122. The processing unit 13 includes an agent determination unit 130, a character setting unit 131, a control unit 132, a moderator judgment unit 133, a viewpoint acquisition unit 134, an idea acquisition unit 135, an idea criticism unit 136, an idea selection unit 137, an analysis unit 138, and a score acquisition unit 139. The output unit 14 includes an answer output unit 141, an idea output unit 142, a score output unit 143, and an error output unit 144.

[0065] The terminal device 2 includes a terminal storage unit 21, a terminal reception unit 22, a terminal processing unit 23, a terminal transmission unit 24, a terminal reception unit 25, and a terminal output unit 26.

[0066] Various types of information are stored in the storage unit 11 constituting the creation support device 1. The various types of information are, for example, transition information (to be described later) and prompt templates (to be described later).

[0067] The transition management unit 111 stores transition information and transition conditions. The transition information is information that specifies the transition between two or more phases for creating an idea. A phase may be called a procedure, a step, a stage, etc. For example, a phase is represented by a node, and a transition between phases is represented by an edge. The transition information usually has two or more phase identifiers. For example, the transition information has edge information that links two phases. For example, the edge information has two phase identifiers. For example, the edge information has an edge identifier.

[0068] A transition condition is a condition under which a phase transitions. A transition condition is a condition associated with a phase before a transition and a phase after the transition. A transition condition is a condition associated with edge information. Two or more transition conditions may be associated with one piece of phase information. When two or more transition conditions are associated with one piece of phase information, there will be two or more phases after the transition. In other words, when two or more transition conditions are associated with one piece of phase information, the phase branches.

[0069] Each of the two or more phases specified by the transition information is a step leading up to the creation of an idea. Each of the two or more phases specified by the transition information includes (a) an idea generation phase and (b) an idea selection phase. Each of the two or more phases specified by the transition information is, for example, (1) an idea perspective / evaluation perspective consideration phase, (2) an idea generation phase, (3) an idea critique phase, and (4) an idea selection phase. Phase identifiers included in the transition information are, for example, "idea perspective / evaluation perspective consideration," "idea generation," "idea critique," and "idea selection."

[0070] The idea selection phase may be a phase in which some ideas are selected from two or more ideas, or a phase in which two or more ideas are ranked. In other words, ranking of ideas may be considered to be included in the selection of ideas.

[0071] The prompt management unit 112 stores various prompt templates. A prompt template is information for constructing a prompt. A prompt template typically includes variables. A prompt is constructed by substituting information into the variable locations. However, a prompt template may also be a prompt that does not include variables. A prompt is information provided to a generating AI. By providing a prompt to a generating AI, an answer is usually passed from the generating AI. A prompt may also be called an inquiry, query, etc. Providing a prompt to a generating AI means, for example, providing the prompt as an argument to the generating AI's API and executing the API.

[0072] The prompt management unit 112 stores one or more setting prompt templates and one or more phase prompt templates. The prompt management unit 112 stores, for example, one or more agent determination prompt templates. The prompt management unit 112 stores, for example, an idea analysis prompt template.

[0073] It is preferable that a setting prompt template for the facilitator and a phase prompt template for the facilitator are stored in the prompt manager 112. The facilitator is a person who acts as a moderator in the role-play for generating ideas.

[0074] A setting prompt template is a prompt template for setting the personality of an agent. One setting prompt template is information for setting the personality of one or more agents. A setting prompt template is associated with, for example, an agent identifier. A setting prompt template has, for example, variables "name", "occupation", "role", and "expertise".

[0075] In this case, an agent is the role played by generative AI. An agent is a virtual person who creates ideas, critiques ideas, and selects ideas. An agent can also be called a persona.

[0076] An agent determination prompt template is a prompt template for determining one or more agents. An agent determination prompt template typically has a variable into which a theme is assigned. An agent determination prompt template has variables into which, for example, one or more idea perspectives and one or more evaluation perspectives are assigned. An example of an agent determination prompt template is, "When generating ideas related to the theme '<theme>', what kind of people are needed?"

[0077] A phase prompt template is a prompt template for one of the phases. A phase prompt template is usually associated with a phase identifier. Note that being associated with a phase identifier may mean that the prompt template identifier includes all or part of the phase identifier.

[0078] The prompt management unit 112 stores, for example, an idea perspective prompt template and an evaluation perspective prompt template in association with the phase identifier "idea perspective / evaluation perspective examination phase." The idea perspective prompt template is a prompt template for acquiring one or more idea perspectives. The idea perspective prompt template typically has a variable into which a theme is assigned. The idea perspective prompt template may, for example, have a variable into which an idea perspective provided by a user is assigned. An example of the idea perspective prompt template is "Please tell me your perspective on ideas related to the theme <theme>!" The evaluation perspective prompt template is a prompt template for acquiring one or more evaluation perspectives. The evaluation perspective prompt template typically has a variable into which a theme is assigned. The evaluation perspective prompt template may, for example, have a variable into which an evaluation perspective provided by a user is assigned. An example of the evaluation perspective prompt template is "Please tell me your perspective on ideas related to the theme <theme>!"

[0079] An idea perspective is a perspective related to an idea. Examples of idea perspectives include "clarification of purpose," "technical feasibility," "legal regulations," and "economic feasibility." An evaluation perspective is a perspective used when evaluating an idea. Examples of evaluation perspectives include "development cost," "usability," "market applicability," and "compliance with regulations."

[0080] The prompt management unit 112 stores, for example, an idea generation prompt template associated with a phase identifier "idea generation phase." An idea generation prompt template is information for constructing a prompt for idea generation. An idea generation prompt template typically has a variable into which a theme is assigned. It is preferable that an idea generation prompt template has variables into which an idea perspective and an evaluation perspective are assigned. It is preferable that an idea generation prompt template has variables into which all or part of the agent information (e.g., "name") is assigned. An example of an idea generation prompt template is, "You are <name>. For <theme>, please come up with an idea while considering the following idea perspectives and the following idea evaluation perspectives. <idea perspective>, <evaluation perspective>."

[0081] The prompt management unit 112 stores, for example, an idea critique prompt template associated with the phase identifier "idea critique phase." An idea critique prompt template is a prompt template for obtaining one or more critiques of an idea. An idea critique prompt template typically has variables into which ideas are assigned. An example of an idea critique prompt template is "Please list the challenges and problems of the idea "<idea>" in relation to <theme>."

[0082] The prompt management unit 112 stores, for example, idea selection prompt templates associated with the phase identifier "idea selection phase." An idea selection prompt template is a prompt template for selecting and ranking ideas. The idea selection prompt template has variables to which two or more ideas are assigned. Note that idea selection can also be done by ranking ideas. Idea selection prompt templates are, for example, "This is a request to the moderator. Please rank the following set of ideas. <Set of ideas>" or "This is a request to the moderator. Please select five ideas that you think are good from the following set of ideas. <Set of ideas>."

[0083] An idea analysis prompt template is a template for a prompt to analyze an idea. An idea analysis prompt template usually has a variable into which an idea is assigned. For example, an idea analysis prompt template has a variable into which a theme is assigned. An example of an idea analysis prompt template is, "Please answer whether the following idea is appropriate as an idea related to <theme> by either "1" for appropriate or "0" for inappropriate! <idea>."

[0084] The attribute value storage unit 113 stores one or more idea attribute values. The idea attribute value is information about the idea to be created. The one or more idea attribute values ​​are, for example, a theme, a purpose, a viewpoint to be included in the idea, and a viewpoint for criticism of the idea. The idea attribute value in the attribute value storage unit 113 may be information input by a user, information stored in advance, or information acquired based on a response from the generation AI.

[0085] The current management unit 114 stores a current identifier. The current identifier is information that identifies the current. The current is the current phase among two or more phases. The two or more phases usually include an idea generation phase and an idea selection phase.

[0086] The receiving unit 12 receives various instructions and information, such as one or more idea attribute values, one or more ideas, and a start instruction.

[0087] Here, acceptance typically refers to the reception of information transmitted via a wired or wireless communication line, but may also be a concept that includes the reception of information input from an input device such as a keyboard, mouse, or touch panel, or the reception of information read from a recording medium such as an optical disk, magnetic disk, or semiconductor memory.

[0088] The start instruction is an instruction to start creating an idea. The start instruction includes one or more idea attribute values, including, for example, a theme.

[0089] The attribute value receiving unit 121 receives one or more idea attribute values. The one or more idea attribute values ​​preferably include a theme. The one or more idea attribute values ​​include, for example, an idea perspective and an evaluation perspective. The attribute value receiving unit 121 receives the one or more idea attribute values ​​from, for example, the terminal device 2.

[0090] The idea receiving unit 122 receives one or more ideas from a person. The idea receiving unit 122 receives one or more ideas from the terminal device 2, for example.

[0091] The processing unit 13 performs various types of processing. The various types of processing are, for example, processing performed by an agent determination unit 130, a character setting unit 131, a control unit 132, a moderator judgment unit 133, a viewpoint acquisition unit 134, an idea acquisition unit 135, an idea criticism unit 136, an idea selection unit 137, an analysis unit 138, or a score acquisition unit 139.

[0092] The agent determination unit 130 determines one or more agents. Determining an agent usually involves obtaining agent information. The agent information may be provided in advance. In other words, the agent information may be stored in the storage unit 11 in advance.

[0093] Agent information is information about an agent. Agent information typically includes an agent identifier. An agent identifier is information that identifies an agent. An agent identifier is, for example, an occupation or specialty. Agent information includes, for example, ability information that indicates the agent's abilities.

[0094] The agent determination unit 130, for example, gives an agent determination prompt to the generated AI and acquires one or more pieces of agent information from the generated AI.

[0095] The agent determination unit 130, for example, assigns a theme or the like to a variable in the agent determination prompt template of the prompt management unit 112, obtains an agent determination prompt, provides the agent determination prompt to the generation AI, and obtains one or more pieces of agent information from the generation AI.

[0096] It is preferable that the agent determination unit 130 acquires one or more pieces of agent information from the generated AI only when the agent acquisition conditions are met. The agent acquisition conditions are conditions for acquiring an agent. For example, the agent acquisition conditions are conditions related to the given agent information. For example, the agent acquisition condition is that the number of pieces of given agent information is 0 or less than a threshold value.

[0097] The agent determination unit 130, for example, acquires one or more pieces of agent information paired with the theme of idea creation from an agent correspondence table. The agent correspondence table is a table showing the correspondence between the theme and agent information. The agent correspondence table has two or more pieces of agent correspondence information. The agent correspondence information has a theme and agent information.

[0098] The person setting unit 131 obtains one or more setting prompts using a setting prompt template, and provides each of the one or more setting prompts to the generation AI.

[0099] The character setting unit 131, for example, obtains a setting prompt template from the prompt management unit 112, replaces variables in the setting prompt template with information about one or more agents to construct a setting prompt, and provides the setting prompt to the generation AI. The character setting unit 131 may construct a setting prompt for each agent and provide the setting prompt for each agent to the generation AI. Alternatively, the character setting unit 131 may construct a setting prompt including agent information for two or more agents and provide the setting prompt to the generation AI. Such agent information may be agent information about one or more agents determined by the agent determination unit 130, or may be agent information provided in advance. In other words, the character setting unit 131 may obtain setting prompts for one or more agents determined by the agent determination unit 130 and provide the setting prompt to the generation AI.

[0100] It is preferable that the character setting unit 131 also provides the generation AI with a setting prompt for the moderator. That is, there may be two or more agents, and one of the two or more agents may be the moderator's agent. The setting prompt for the moderator is stored, for example, in the prompt management unit 112.

[0101] The control unit 132 determines whether the current phase matches a transition condition associated with the current phase, and if the transition condition is met, changes the current phase from the current phase to a post-transition phase associated with the transition condition. The control unit 132 usually changes the current phase identifier of the current management unit 114 to the phase identifier of the post-transition phase.

[0102] The moderator determination unit 133 determines whether the prompt given to the generation AI or the response from the generation AI satisfies the moderator's statement condition. The moderator's statement condition may be any transition condition.

[0103] The moderator's speaking conditions are the conditions under which the moderator's agent will speak, and are conditions related to the prompt given to the generation AI or the response from the generation AI. Examples of moderator's speaking conditions include "ideas have been obtained from all two or more parties" and "a certain amount of time has passed without any human input."

[0104] The viewpoint acquisition unit 134 acquires one or more idea viewpoints or one or more evaluation viewpoints. The viewpoint acquisition unit 134 acquires, for example, one or more idea viewpoints accepted by the attribute value accepting unit 121. The viewpoint acquisition unit 134 acquires, for example, one or more evaluation viewpoints accepted by the attribute value accepting unit 121.

[0105] The viewpoint acquisition unit 134, for example, assigns a theme to a variable of an idea viewpoint prompt template, constructs an idea viewpoint prompt, passes the idea viewpoint prompt to the generation AI, and acquires one or more idea viewpoints from the generation AI. Note that an idea viewpoint prompt template is, for example, "Please tell me the viewpoint when creating an idea on the following theme! <theme>". Furthermore, <theme> is a variable to which the acquired theme is assigned. A character string enclosed in "<" and ">" indicates a variable.

[0106] The viewpoint acquisition unit 134, for example, constructs an evaluation viewpoint prompt by substituting a theme for a variable of an evaluation viewpoint prompt template, passes the evaluation viewpoint prompt to the generation AI, and acquires one or more evaluation viewpoints from the generation AI. Note that an evaluation viewpoint prompt template is, for example, "Please tell me the viewpoint for evaluating ideas on the following theme! <theme>".

[0107] The viewpoint acquisition unit 134 preferably determines, for example, whether viewpoint acquisition conditions are met, and acquires a viewpoint only when it is determined that the conditions are met. The viewpoint acquisition conditions are, for example, conditions related to idea viewpoints. The viewpoint acquisition conditions are, for example, conditions related to evaluation viewpoints. The viewpoint acquisition unit 134 determines to acquire an idea viewpoint, for example, when the number of accepted idea viewpoints is equal to or less than a threshold. The viewpoint acquisition unit 134 determines to acquire an evaluation viewpoint, for example, when the number of accepted evaluation viewpoints is equal to or less than a threshold. The threshold is a natural number equal to or greater than 1.

[0108] The idea acquisition unit 135 acquires a prompt using a phase prompt template corresponding to each of one or more phases, and gives the prompt to the generation AI to acquire two or more ideas.

[0109] The idea acquisition unit 135, for example, acquires a prompt using a phase prompt template corresponding to the idea generation phase and one or more idea attribute values, and provides the prompt to one or more agents of the generation AI to acquire two or more ideas. Here, the one or more idea attribute values ​​are, for example, one or more pieces of information selected from a theme, an idea perspective, and an evaluation perspective. It is preferable that the idea acquisition unit 135 acquires a prompt using a theme and an idea perspective, and provides the prompt to one or more agents of the generation AI to acquire two or more ideas. The idea acquisition unit 135 may also acquire a prompt using a theme, an idea perspective, and an evaluation perspective, and provide the prompt to one or more agents of the generation AI to acquire two or more ideas.

[0110] Providing a prompt to one or more agents of the generated AI means, for example, providing the generated AI with a prompt that includes a character string indicating that the prompt is a query for one or more agents. Providing a prompt to the generated AI usually means executing the generated AI's API, which includes the prompt as an argument.

[0111] The idea acquisition unit 135 corresponds to, for example, the idea generation phase, acquires a prompt using a phase prompt template for the moderator and one or more idea attribute values, provides the prompt to the generation AI, and acquires ideas from one or more of the two or more agents.

[0112] For example, when the moderator judgment unit 133 judges that the moderator statement conditions are met, the idea acquisition unit 135 corresponds to the idea generation phase, acquires a prompt for each of one or more agents other than the moderator, gives the prompt to the generation AI, and acquires two or more ideas. It is preferable that the idea acquisition unit 135 acquires an idea from each of one or more agents not just once, but twice or more times.

[0113] The idea critique unit 136 acquires one or more critiques for each of the two or more ideas acquired by the idea acquisition unit 135. The critiques may be called problems, issues, etc. The critiques are character strings.

[0114] The idea critique unit 136, for example, constructs a critique prompt by substituting the idea acquired by the idea acquisition unit 135 into a variable of a critique prompt template corresponding to the idea critique phase, passes the critique prompt to the generation AI, and acquires one or more critiques from the generation AI. The idea critique unit 136 may also substituting one or more evaluation perspectives into a variable of the critique prompt template, constructs a critique prompt, passes the critique prompt to the generation AI, and acquires one or more critiques from the generation AI.

[0115] The idea critique unit 136, for example, for each of two or more ideas acquired by the idea acquisition unit 135, assigns the idea to a variable in a critique prompt template, constructs a critique prompt, passes the critique prompt to the generation AI, and acquires one or more critiques from the generation AI.

[0116] The idea critique unit 136, for example, for each of two or more ideas and each of two or more agents acquired by the idea acquisition unit 135, assigns some or all of the idea and agent information to the variables of a critique prompt template to construct a critique prompt, passes the critique prompt to the generation AI, and acquires one or more critiques from the generation AI.

[0117] The idea critique unit 136 may, for example, construct a critique prompt for each of two or more evaluation perspectives, pass the critique prompt to the generation AI, and obtain one or more critiques from the generation AI. A critique prompt template in such a case may be, for example, "You are <name>. Regarding <theme>, please list the issues and problems with idea "<idea>" from the evaluation perspective "<evaluation perspective>." Through the above processing, the idea critique unit 136 obtains one or more critiques for each of two or more ideas and each of two or more evaluation perspectives.

[0118] The idea selection unit 137 acquires a prompt using a phase prompt template corresponding to the idea selection phase and the two or more ideas acquired by the idea acquisition unit 135, provides the prompt to the generation AI, and acquires one or more ideas selected based on the two or more ideas acquired by the idea acquisition unit 135. Note that the one or more selected ideas may be two or more ranked ideas.

[0119] For example, the idea selection unit 137 constructs an idea selection prompt by substituting a set of two or more ideas and some or all of the agent information into variables of an idea selection prompt template for each of two or more agents, and provides the idea selection prompt to a generation AI to obtain one or two or more ideas.

[0120] The idea selection unit 137, for example, assigns two or more ideas acquired by the idea acquisition unit 135 to variables of a phase prompt template corresponding to the idea selection phase, acquires a prompt, provides the prompt to the generation AI, and acquires one or more selected ideas.

[0121] The idea selection unit 137, for example, assigns two or more ideas acquired by the idea acquisition unit 135 and one or more critiques acquired by the idea critique unit 136 to variables of a phase prompt template corresponding to the idea selection phase, acquires a prompt, provides the prompt to the generation AI, and acquires one or more selected ideas.

[0122] The idea selection unit 137 acquires a prompt using a phase prompt template corresponding to the idea selection phase, two or more ideas acquired by the idea acquisition unit 135, and one or more ideas accepted by the idea acceptance unit 122, and provides the prompt to the generation AI to acquire one or more ideas selected based on the two or more ideas acquired by the idea acquisition unit 135 and one or more ideas accepted by the idea acceptance unit 122.

[0123] The idea selection unit 137 acquires a prompt using, for example, a phase prompt template corresponding to the idea selection phase and two or more ideas excluding ideas that the analysis unit 138 has determined do not satisfy the consideration conditions, provides the prompt to the generation AI, and acquires one or more ideas selected based on the two or more ideas acquired by the idea acquisition unit 135 and one or more ideas accepted by the idea acceptance unit 122.

[0124] The idea selection unit 137 corresponds to, for example, the idea selection phase, acquires a prompt using a phase prompt template for the moderator and two or more ideas acquired by the idea acquisition unit 135, provides the prompt to the generation AI, and acquires one or more ideas selected based on the two or more ideas acquired by the idea acquisition unit 135.

[0125] The analysis unit 138 analyzes each of the one or more ideas received by the idea receiving unit 122 and acquires the error class. Note that the two or more error classes usually include "normal."

[0126] The analysis unit 138 may construct an idea analysis prompt by substituting one or more ideas received by the idea receiving unit 122 into variables of an idea analysis prompt template, pass the idea analysis prompt to the generation AI, and obtain from the generation AI an answer as to whether the idea is normal or an error.

[0127] The analysis unit 138 may, for example, perform natural language processing on character strings that are one or more ideas received by the idea receiving unit 122, determine whether or not the character strings correspond to "ideas," and acquire a class. Note that the technology for classifying character strings is a well-known technology.

[0128] The score acquisition unit 139 performs natural language processing on the two or more ideas acquired by the idea acquisition unit 135 or the idea selection unit 137, and acquires a score related to the diversity of the ideas. For example, the score acquisition unit 139 uses character strings of the two or more ideas to acquire a Remote-Clique (RC value) or a Chamfer Distance (CD value), which is a score related to the diversity of the two or more ideas. The technology for acquiring the RC value and the CD value is well known, so a detailed description thereof will be omitted.

[0129] The score acquiring unit 139 may acquire a score for each of one or more ideas acquired by the idea acquiring unit 135 or the idea selecting unit 137. It is preferable that the score acquiring unit 139 acquires a score for each idea using, for example, one or more critiques corresponding to the idea, acquired by the idea critique unit 136. It does not matter how the score acquiring unit 139 acquires the score for each idea using the one or more critiques for each idea.

[0130] For example, the score acquiring unit 139 constructs a ranking prompt by substituting two or more ideas and one or more comments for each of the two or more evaluation criteria into variables of a ranking prompt template for prompting a user to rank two or more ideas, and provides the ranking prompt to the generation AI, which then obtains a ranking of each of the two or more ideas for each of the two or more evaluation criteria from the generation AI. Next, the score acquiring unit 139 acquires a higher score for each of the two or more ideas, the higher the ranking for each of the two or more evaluation criteria. Note that, for example, the score acquiring unit 139 acquires a score for each of the two or more evaluation criteria for each of the two or more ideas, and calculates the score using an increasing function (e.g., sum, average, median, etc.) with the two or more scores as a parameter. Note that the score acquiring unit 139 calculates the score for each of the two or more evaluation criteria for an idea using, for example, a decreasing function with the ranking of the evaluation criteria as a parameter. The decreasing function is, for example, "score for each evaluation criteria = (100 * (number of ideas - ranking)) / (number of evaluation criteria x (number of ideas - 1))".

[0131] Furthermore, for example, the greater the number of reviews corresponding to an idea, the lower the score acquired by the score acquiring unit 139. For example, the score acquiring unit 139 calculates the score using a decreasing function that uses the number of reviews corresponding to the idea as a parameter.

[0132] When the score acquisition unit 139 acquires scores of 2 or more, the idea selection unit 137 may use the scores to associate (rank) the two or more ideas with rankings based on the scores. The idea selection unit 137 assigns a higher ranking to an idea with a higher score.

[0133] The output unit 14 outputs various types of information, such as all or part of the answer passed from the generation AI, or one or more ideas.

[0134] It is preferable that the output unit 14 outputs all or part of the answer passed from the generation AI in a manner that allows the current identified by the current identifier to be grasped.

[0135] Here, output usually means transmission to terminal device 2, but it may also be a concept that includes display on a display, projection using a projector, printing on a printer, sound output, storage on a recording medium, and handing over the processing results to other processing devices or other programs.

[0136] The answer output unit 141 outputs all or part of the answer passed from the generation AI. The answer passed from the generation AI may include an idea.

[0137] The idea output unit 142 outputs one or more ideas acquired by the idea selection unit 137. It is preferable that the idea output unit 142 outputs two or more ideas ranked by the idea selection unit 137 in a manner that makes the ranking clear.

[0138] The score output unit 143 outputs the score acquired by the score acquisition unit 139 .

[0139] The error output unit 144 outputs an error message when the analysis unit 138 determines that the review conditions are not satisfied. It is preferable that the error output unit 144 outputs an error message including the class of the error.

[0140] Various types of information are stored in the terminal storage unit 21 included in the terminal device 2. The various types of information are, for example, user identifiers. A user identifier is information that identifies a user.

[0141] The terminal reception unit 22 receives various types of information, instructions, etc. The various types of information, instructions, etc. may be, for example, one or more ideas, one or more idea attribute values, etc. The means for inputting the various types of information, instructions, etc. may be any means, such as a touch panel, keyboard, mouse, microphone, or menu screen.

[0142] The terminal processing unit 23 performs various types of processing. For example, the various types of processing are processing to convert received information, instructions, etc. into information, instructions, etc. with a structure to be transmitted. For example, the various types of processing are processing to convert received information into information with a structure to be output.

[0143] The terminal transmitting unit 24 transmits various types of information, instructions, etc. to the creation support device 1. The various types of information, instructions, etc. are, for example, one or more ideas and one or more idea attribute values.

[0144] The terminal receiving unit 25 receives various types of information, such as answers, ideas, scores, and error messages.

[0145] The terminal output unit 26 outputs various types of information, such as answers, ideas, scores, and error messages.

[0146] The storage unit 11, transition management unit 111, prompt management unit 112, attribute value storage unit 113, current management unit 114, and terminal storage unit 21 are preferably non-volatile recording media, but can also be realized as volatile recording media.

[0147] There is no restriction on the process by which information is stored in the storage unit 11 etc. For example, information may be stored in the storage unit 11 etc. via a recording medium, information transmitted via a communication line etc. may be stored in the storage unit 11 etc., or information input via an input device may be stored in the storage unit 11 etc.

[0148] The reception unit 12, the attribute value reception unit 121, and the idea reception unit 122 are usually realized by wireless or wired communication means, but may also be realized by a means for receiving broadcasts, a device driver for an input means such as a touch panel or keyboard, or control software for a menu screen.

[0149] The processing unit 13, agent determination unit 130, character setting unit 131, control unit 132, moderator judgment unit 133, viewpoint acquisition unit 134, idea acquisition unit 135, idea criticism unit 136, idea selection unit 137, analysis unit 138, score acquisition unit 139, and device processing unit 23 can typically be realized by a processor, memory, etc. The processing procedures of the processing unit 13, etc. are typically realized by software, and the software is recorded on a recording medium such as a ROM. However, they may also be realized by hardware (dedicated circuit). The processor may be a CPU, MPU, GPU, etc., and the type is not important.

[0150] The output unit 14, answer output unit 141, idea output unit 142, score output unit 143, and error output unit 144 are usually realized by wireless or wired communication means, but may also be realized by driver software for an output device such as a display or speaker, or by a combination of driver software for an output device and the output device.

[0151] The terminal reception unit 22 can be realized by a device driver for an input means such as a touch panel or a keyboard, or control software for a menu screen.

[0152] The terminal transmitting unit 24 is usually realized by a wireless or wired communication means, but may also be realized by a broadcasting means.

[0153] The terminal receiving unit 25 is usually realized by a wireless or wired communication means, but may also be realized by a means for receiving broadcasts.

[0154] The terminal output unit 26 may or may not include an output device such as a display, a speaker, etc. The terminal output unit 26 may be realized by driver software for an output device, or by a combination of driver software for an output device and the output device, etc.

[0155] Next, an example of the operation of the creation support device 1 will be described with reference to the flowchart of FIG.

[0156] (Step S401) The attribute value receiving unit 121 determines whether or not a start instruction has been received. If a start instruction has been received, the process proceeds to step S402, and if not, the process returns to step S401.

[0157] (Step S402) The processing unit 13 performs pre-processing. An example of the pre-processing will be described with reference to the flowchart of FIG.

[0158] (Step S403) The viewpoint acquisition unit 134 judges whether or not to acquire a viewpoint. If a viewpoint is to be acquired, the process proceeds to step S404, and if not, the process proceeds to step S405.

[0159] (Step S404) The viewpoint acquisition unit 134 acquires a viewpoint. An example of such viewpoint acquisition processing will be described with reference to the flowchart of FIG.

[0160] (Step S405) The idea acquisition unit 135 performs an idea acquisition process. An example of the idea acquisition process will be described with reference to the flowchart in FIG.

[0161] (Step S406) The control unit 132 acquires a transition condition for moving from the idea generation phase to the next phase from the transition management unit 111. The control unit 132 determines whether the transition condition is satisfied. If the transition condition is satisfied, the process proceeds to step S407, and if the transition condition is not satisfied, the process returns to step S405.

[0162] (Step S407) The control unit 132 refers to the transition management unit 111 and determines whether the next phase is an idea criticism phase. If it is an idea criticism phase, the process proceeds to step S408; if it is not an idea criticism phase, the process proceeds to step S410. Here, the control unit 132 stores the phase identifier of the next phase in the current management unit 114.

[0163] (Step S408) The idea critique unit 136 acquires the critique of the idea. An example of such idea critique processing will be described with reference to the flowchart of FIG.

[0164] (Step S409) The control unit 132 acquires the transition condition for moving from the idea criticism phase to the next phase from the transition management unit 111, and determines whether the transition condition is satisfied. If the transition condition is satisfied, the process proceeds to step S410, and if the transition condition is not satisfied, the process returns to step S408.

[0165] (Step S410) The idea selection unit 137 performs an idea selection process. An example of the idea selection process will be described with reference to the flowchart in Fig. 12. Here, the control unit 132 stores the phase identifier of the next phase, "idea selection," in the current management unit 114.

[0166] (Step S411) The idea output unit 142 outputs one or more ideas acquired in step S410. Return to step S401. Note that the idea output unit 142 transmits one or more ideas to the terminal device 2, for example.

[0167] (Step S412) The score acquisition unit 139 acquires a score based on one or more ideas acquired in step S410.

[0168] (Step S413) The score output unit 143 outputs the score acquired in step S412.

[0169] In the flowchart of FIG. 4, the answer output unit 141 may output all or part of the answer acquired from the generation AI.

[0170] In the flowchart of FIG. 4, the process ends when the power is turned off or an interrupt occurs to end the process.

[0171] Next, an example of the pre-processing in step S402 will be described with reference to the flowchart in FIG.

[0172] (Step S501) The processing unit 13 acquires the theme etc. included in the start instruction. The theme etc. is, for example, a theme and one or more pieces of agent information, or a theme. The theme etc. includes, for example, one or more idea perspectives and one or more evaluation perspectives.

[0173] (Step S502) The processing unit 13 acquires one or more idea perspectives included in the start instruction. Note that the processing unit 13 does not necessarily need to acquire idea perspectives here.

[0174] (Step S503) The processing unit 13 acquires one or more evaluation aspects included in the start instruction. Note that the processing unit 13 does not necessarily need to acquire any evaluation aspects.

[0175] (Step S504) The agent determination unit 130 determines whether or not agent information is present in the information acquired in step S501. If agent information is present, the process proceeds to step S505, and if not, the process proceeds to step S506.

[0176] (Step S505) The agent determining unit 130 acquires one or more pieces of agent information from the information acquired in step S501. Then, the process proceeds to step S507.

[0177] (Step S506) The agent determination unit 130 performs an agent determination process. An example of the agent determination process will be described with reference to the flowchart in FIG.

[0178] (Step S507) The person setting unit 131 performs a person setting process and returns to the upper level process. An example of the person setting process will be described with reference to the flowchart in FIG.

[0179] 5, if the start instruction does not have agent information, the agent determination unit 130 acquires agent information through the agent determination process. However, in the flowchart of FIG. 5, if the amount of agent information included in the start instruction is equal to or less than a threshold, the agent determination unit 130 may acquire agent information through the agent determination process.

[0180] Next, an example of the agent determination process in step S506 will be described with reference to the flowchart in FIG.

[0181] (Step S601) The agent determination unit 130 acquires an agent determination prompt template from the prompt management unit 112.

[0182] (Step S602) The agent determining unit 130 acquires the accepted theme and the like.

[0183] (Step S603) The agent determination unit 130 assigns the theme and the like acquired in step S602 to the variables of the agent determination prompt template acquired in step S601, and acquires an agent determination prompt.

[0184] (Step S604) The agent determination unit 130 passes the agent determination prompt acquired in step S603 to the generation AI.

[0185] (Step S605) The agent determining unit 130 judges whether or not an answer has been acquired from the generating AI. If an answer has been acquired, the process proceeds to step S606, and if an answer has not been acquired, the process returns to step S605.

[0186] (Step S606) The agent determination unit 130 acquires one or more pieces of agent information from the response.

[0187] (Step S607) The response output unit 141 outputs one or more pieces of agent information. The process returns to the upper level process. The response output unit 141 transmits one or more pieces of agent information to the terminal device 2, for example.

[0188] Next, an example of the person setting process in step S507 will be described with reference to the flowchart in FIG.

[0189] (Step S701) The person setting unit 131 acquires a setting prompt template from the prompt management unit 112.

[0190] (Step S702) The person setting unit 131 acquires the agent information of the facilitator. Note that the agent information of the facilitator may be included in the start instruction or may be stored in the storage unit 11 in advance.

[0191] (Step S703) The person setting unit 131 assigns the agent information of the facilitator acquired in step S702 to the variable of the setting prompt template acquired in step S701, and acquires the setting prompt of the facilitator.

[0192] (Step S704) The person setting unit 131 gives the setting prompt of the facilitator acquired in step S703 to the generation AI.

[0193] (Step S705) The person setting unit 131 assigns 1 to a counter i.

[0194] (Step S706) The person setting unit 131 determines whether or not the i-th agent information, excluding the agent information of the facilitator, exists among the acquired agent information. If the i-th agent information exists, the process proceeds to step S707; if not, the process returns to the upper processing.

[0195] (Step S707) The person setting unit 131 acquires the i-th agent information.

[0196] (Step S708) The person setting unit 131 assigns the i-th agent information to the variable of the setting prompt template, and acquires the setting prompt of the i-th agent.

[0197] (Step S709) The character setting unit 131 gives the setting prompt of the ith agent acquired in step S708 to the generation AI.

[0198] (Step S710) The person setting unit 131 increments the counter i by 1. The process returns to step S706.

[0199] Next, an example of the viewpoint acquisition process in step S404 will be described with reference to the flowchart in FIG.

[0200] (Step S801) The viewpoint acquiring unit 134 acquires an idea viewpoint prompt template from the prompt managing unit 112.

[0201] (Step S802) The viewpoint acquiring unit 134 acquires a theme or the like, which is information to be substituted into the idea viewpoint prompt template acquired in step S801. The viewpoint acquiring unit 134 assigns the theme or the like to the idea viewpoint prompt template and acquires an idea viewpoint prompt.

[0202] (Step S803) The perspective acquisition unit 134 provides the idea perspective prompt acquired in step S802 to the generation AI.

[0203] (Step S804) The viewpoint acquisition unit 134 determines whether or not an answer has been acquired from the generation AI. If an answer has been acquired, the process proceeds to step S805, and if an answer has not been acquired, the process returns to step S804.

[0204] (Step S805) The viewpoint acquisition unit 134 acquires one or more idea viewpoints from the answer acquired from the generation AI in step S804.

[0205] (Step S806) The answer output unit 141 outputs one or more idea perspectives acquired in step S805. The answer output unit 141 transmits one or more idea perspectives to the terminal device 2, for example.

[0206] (Step S807) The viewpoint acquiring unit 134 acquires an evaluation viewpoint prompt template from the prompt managing unit 112.

[0207] (Step S808) The viewpoint acquiring unit 134 acquires a theme or the like, which is information to be substituted into the evaluation viewpoint prompt template acquired in step S807. The viewpoint acquiring unit 134 assigns the theme or the like to the evaluation viewpoint prompt template and acquires an evaluation viewpoint prompt.

[0208] (Step S809) The viewpoint acquisition unit 134 provides the evaluation viewpoint prompt acquired in step S808 to the generation AI.

[0209] (Step S810) The viewpoint acquisition unit 134 determines whether or not an answer has been acquired from the generation AI. If an answer has been acquired, the process proceeds to step S811, and if an answer has not been acquired, the process returns to step S810.

[0210] (Step S811) The viewpoint acquisition unit 134 acquires one or more evaluation viewpoints from the answer acquired from the generation AI in step S810.

[0211] (Step S812) The answer output unit 141 outputs the one or more evaluation criteria acquired in step S811. The answer output unit 141 transmits the one or more evaluation criteria to the terminal device 2, for example.

[0212] Next, an example of the idea acquisition process in step S405 will be described with reference to the flowchart in FIG.

[0213] (Step S901) The idea acquiring unit 135 acquires an idea generation prompt template from the prompt managing unit 112.

[0214] (Step S902) The idea obtaining unit 135 assigns 1 to a counter i.

[0215] (Step S903) The idea acquisition unit 135 determines whether or not the i-th agent excluding the moderator exists. If the i-th agent exists, the process proceeds to step S904, and if the i-th agent does not exist, the process proceeds to step S911. Note that whether or not the i-th agent exists depends on whether or not the i-th agent information exists.

[0216] (Step S904) The idea acquiring unit 135 acquires the ith agent information.

[0217] (Step S905) The idea acquiring unit 135 assigns part or all of the ith agent information to the variables of the idea generating prompt template, and acquires an idea generating prompt for the ith agent.

[0218] (Step S906) The idea acquisition unit 135 gives the idea generation prompt acquired in step S905 to the generation AI.

[0219] (Step S907) The idea acquisition unit 135 determines whether or not an answer has been acquired from the generation AI. If an answer has been acquired, the process proceeds to step S908, and if an answer has not been acquired, the process returns to step S907.

[0220] (Step S908) The idea acquiring unit 135 acquires one or more ideas from the answers acquired in step S907.

[0221] (Step S909) The idea output unit 142 outputs one or more ideas acquired in step S908. The idea output unit 142 transmits the one or more ideas to the terminal device 2, for example.

[0222] (Step S910) The idea acquisition unit 135 increments the counter i by 1. The process returns to step S903.

[0223] (Step S911) The idea acquiring unit 135 determines whether to accept an input of an idea from a person. If an input of an idea from a person is accepted, the process proceeds to step S912, and if not, the process returns to the upper processing.

[0224] (Step S912) The idea obtaining unit 135 assigns 1 to a counter i.

[0225] (Step S913) The idea acquisition unit 135 determines whether or not there is an i-th person who will input an idea. If there is an i-th person, the process proceeds to step S914, and if there is no i-th person, the process returns to the upper process.

[0226] (Step S914) The idea acquisition unit 135 determines whether or not one or more ideas have been received from a person. If an idea has been received, the process proceeds to step S915, and if not, the process returns to step S914. Note that if a timeout occurs here, the process may proceed to S919. Furthermore, receiving an idea from a person may be, for example, receiving an idea from the terminal device 2.

[0227] (Step S915) The analysis unit 138 analyzes the one or more ideas received in step S914. An example of such idea analysis processing will be described with reference to the flowchart of FIG.

[0228] (Step S916) The idea acquisition unit 135 determines whether the analysis result in step S915 was an "error." If it is an "error," the process proceeds to step S918, and if it is not an "error," the process proceeds to step S917.

[0229] (Step S917) The idea acquiring unit 135 acquires one or more ideas accepted in step S914.

[0230] (Step S918) The error output unit 144 outputs an error message. For example, the error output unit 144 transmits the error message to the terminal device 2 that transmitted the one or more ideas.

[0231] (Step S919) The idea obtaining unit 135 increments the counter i by 1. The process returns to step S913.

[0232] Next, an example of the idea analysis process in step S915 will be described with reference to the flowchart of FIG.

[0233] (Step S1001) The analysis unit 138 acquires an idea analysis prompt template from the prompt management unit 112.

[0234] (Step S1002) The analysis unit 138 assigns 1 to a counter i.

[0235] (Step S1003) The analysis unit 138 determines whether or not the i-th idea exists. If the i-th idea exists, the process proceeds to step S1004, and if not, the process returns to the upper level process.

[0236] (Step S1004) The analysis unit 138 acquires the i-th idea.

[0237] (Step S1005) The analysis unit 138 assigns the ith idea to the idea analysis prompt template acquired in step S1001, and acquires an idea analysis prompt.

[0238] (Step S1006) The analysis unit 138 passes the idea analysis prompt acquired in step S1005 to the generation AI.

[0239] (Step S1007) The analysis unit 138 determines whether or not an answer has been acquired from the generation AI. If an answer has been acquired, the process proceeds to step S1008, and if an answer has not been acquired, the process returns to step S1007.

[0240] (Step S1008) The analysis unit 138 acquires an analysis result (for example, "normal" or "error") from the answer acquired in step S1007, and associates it with the i-th idea.

[0241] (Step S1009) The analysis unit 138 increments the counter i by 1. The process returns to step S1003.

[0242] Next, an example of the idea criticism process in step S408 will be described with reference to the flowchart of FIG.

[0243] (Step S1101) The idea critique unit 136 acquires an idea critique prompt template from the prompt management unit 112.

[0244] (Step S1102) The idea criticism unit 136 assigns 1 to a counter i.

[0245] (Step S1103) The idea criticism unit 136 determines whether or not the i-th idea to be criticized exists. If the i-th idea exists, the process proceeds to step S1104, and if not, the process returns to the upper level process.

[0246] (Step S1104) The idea criticism unit 136 assigns 1 to a counter j.

[0247] (Step S1105) The idea criticism unit 136 determines whether or not the j-th agent exists. If the j-th agent exists, the process proceeds to step S1106, and if not, the process proceeds to step S1113.

[0248] (Step S1106) The idea criticism unit 136 acquires the j-th agent information.

[0249] (Step S1107) The idea critique unit 136 substitutes the ith idea and part or all of the jth agent information into the idea critique prompt template acquired in step S1101 to acquire an idea critique prompt.

[0250] (Step S1108) The idea critique unit 136 passes the idea critique prompt acquired in step S1107 to the generation AI.

[0251] (Step S1109) The idea criticism unit 136 determines whether or not an answer has been obtained from the generating AI. If an answer has been obtained, the process proceeds to step S1110, and if not, the process returns to step S1109.

[0252] (Step S1110) The idea criticism unit 136 acquires one or more criticisms from the answer from the generation AI acquired in step S1109.

[0253] (Step S1111) The answer output unit 141 outputs one or more criticisms acquired in step S1110 in association with the i-th idea.

[0254] (Step S1112) The idea criticism unit 136 increments the counter j by 1. The process returns to step S1105.

[0255] (Step S1113) The idea criticism unit 136 increments the counter i by 1. The process returns to step S1103.

[0256] Next, an example of the idea selection process in step S410 will be described with reference to the flowchart in FIG.

[0257] (Step S1201) The idea selection unit 137 acquires an idea selection prompt template from the prompt management unit 112.

[0258] (Step S1202) The idea selection unit 137 acquires the idea acquired by the idea acquisition unit 135, or the idea acquired by the idea acquisition unit 135 and the idea accepted by the idea acceptance unit 122. It is preferable that the idea selection unit 137 acquires one or more criticisms acquired by the idea criticism unit 136.

[0259] (Step S1203) The idea selection unit 137 assigns two or more ideas, or two or more ideas and one or more reviews, to the idea selection prompt template acquired in step S1201 to acquire an idea selection prompt.

[0260] (Step S1204) The idea selection unit 137 passes the idea selection prompt acquired in step S1203 to the generation AI.

[0261] (Step S1205) The idea selection unit 137 determines whether or not an answer has been acquired from the generation AI. If an answer has been acquired, the process proceeds to step S1206, and if an answer has not been acquired, the process returns to step S1205.

[0262] (Step S1206) The idea selection unit 137 acquires one or more ideas from the answers acquired in step S1205. The process returns to the upper level process. Note that here, it is preferable that the idea selection unit 137 acquires two or more ranked ideas.

[0263] A specific example of the operation of the information system A in this embodiment will be described below.

[0264] Currently, the transition management unit 111 of the creation support device 1 stores a transition management table shown in FIG. 13. The transition management table is a table that manages phase transitions for idea creation. The transition management table manages two or more records each having an "ID," "phase identifier," "transition condition," and "next phase identifier." The "ID" is information that identifies a record. The combination of the "transition condition" and "next phase identifier" indicates that in the current phase identified by the "phase identifier," if the transition condition is met, the system will transition to the next phase indicated by the next phase identifier.

[0265] In FIG. 13, if the number of idea viewpoints and the number of evaluation viewpoints included in the start instruction satisfy the transition condition of "ID=1", the "idea viewpoint / evaluation viewpoint consideration" phase is skipped.

[0266] In the record "ID=2" in Figure 13, two pairs of "transition condition" and "next phase identifier" correspond to one phase identifier "idea generation." In other words, the record "ID=2" indicates that the phase "idea generation" transitions to different next phases depending on the matching transition condition. The record "ID=2" indicates that branching is possible in the phase transition.

[0267] The prompt management unit 112 also stores a prompt management table shown in FIG. 14. The prompt management table is a table for managing prompt templates. The prompt management table manages a large number of records having "prompt template identifiers" and "prompt templates." Each "prompt template identifier" may be associated with a phase identifier. The "idea perspective prompt template" and "evaluation perspective prompt template" are associated with the phase identifier "idea perspective / evaluation consideration." The "idea generation prompt template" is associated with the phase identifier "idea generation." The "idea critique prompt template" is associated with the phase identifier "idea critique." The "idea selection prompt template" is associated with the phase identifier "idea selection." The "idea generation prompt template" has the variable <idea perspective> and the variable <evaluation perspective>, but it does not have to have the variable <evaluation perspective>. The "idea critique prompt template" may have the variable <evaluation perspective>. In such a case, the "idea critique prompt template" might be, for example, "You are <name>. Regarding <topic>, please list the challenges and problems of the idea "<idea>" by taking into consideration the following idea evaluation criteria: [Evaluation criteria] <Evaluation criteria>."

[0268] In the above situation, User U wants to create an idea for a "solution using AI and a camera" together with two or more agents of the generation AI, and creates the following: <Theme> Solution using AI and a camera <Agent><Name> Person A <Agent name> AI Engineer <Role> Utilizing AI-related technologies to promote the use of AI in ideas and products and improve problem-solving capabilities. <Persona> ['Has deep knowledge and experience of AI technologies such as machine learning, deep learning, and natural language processing', 'Skilled in dataset preprocessing, model training and evaluation, and deployment of AI solutions', 'Conscious about developing AI with consideration for its ethical aspects and social impact']< / エージェント> " is input to the terminal device 2.

[0269] Next, the terminal device 2 accepts the start instruction and transmits the start instruction to the creation support device 1.

[0270] Next, the reception unit 12 of the creation support device 1 receives the start instruction. Next, the processing unit 13 performs pre-processing as follows.

[0271] That is, the processing unit 13 acquires the theme "Solution using AI and a camera" included in the start instruction. The processing unit 13 also attempts to acquire idea perspectives and evaluation perspectives from the start instruction, but is unable to do so. Next, the agent determination unit 130 acquires an agent determination prompt template from the prompt management table (FIG. 14). The agent determination unit 130 acquires the theme and "<agent><name>Mr. A <agent name>AI Engineer..." from the start instruction.< / エージェント> ". Then, the agent determination unit 130 assigns "Solution using AI and a camera" to the variable <theme> of the agent determination prompt template, assigns the <agent name> "AI engineer" to the variable <user-specified agent>, and acquires the agent determination prompt. Next, the agent determination unit 130 passes the agent determination prompt to the generation AI. Next, the agent determination unit 130 acquires a response from the generation AI and acquires one or more pieces of agent information from the response. Here, it is assumed that the agent determination unit 130 has acquired the agent information of "data scientist" and "sales" (see FIG. 15). In other words, the agent determination unit 130 has acquired the agent information of "AI engineer" given by the user U and the agent information of "data scientist" and "sales" that are responses from the generation AI.

[0272] Next, the character setting unit 131 performs the character setting process as follows, following the process described using the flowchart in FIG. 7. That is, the character setting unit 131 acquires a setting prompt template from the prompt management table (FIG. 14). The character setting unit 131 substitutes <name><agent name><role><persona> into the setting prompt template from each of the agent information input by the user U and the agent information acquired by the agent determination unit 130, acquires a setting prompt for each person, and passes the setting prompt to the generation AI. As a result, characters for AI engineer, data science, and sales have been set. The character setting unit 131 also acquires agent information for the moderator from the storage unit 11, performs similar processing, and sets the moderator as a character. This completes the pre-processing.

[0273] Next, the perspective acquisition unit 134 determines that the start instruction does not include an idea perspective or an evaluation perspective, and therefore satisfies the idea perspective acquisition condition (number of idea perspectives included in the start instruction <= threshold a) and the evaluation perspective acquisition condition (number of evaluation perspectives included in the start instruction <= threshold b).The perspective acquisition unit 134 then acquires an idea perspective prompt template from the prompt management table (Figure 14).The perspective acquisition unit 134 acquires a <theme> and an <input idea perspective> to be substituted into the idea perspective prompt template. Note that the <input idea perspective> is not acquired here.The perspective acquisition unit 134 assigns the theme to the idea perspective prompt template and acquires an idea perspective prompt.The perspective acquisition unit 134 then provides the idea perspective prompt to the generation AI.The perspective acquisition unit 134 then acquires, for example, idea perspectives such as "entered market" and "technological innovativeness."

[0274] The viewpoint acquisition unit 134 also acquires an evaluation viewpoint prompt template from the prompt management table (Figure 14). The viewpoint acquisition unit 134 acquires a <theme> and an <input evaluation viewpoint> to be substituted into the evaluation viewpoint prompt template. Note that the <input evaluation viewpoint> is not acquired here. The viewpoint acquisition unit 134 assigns the theme to the evaluation viewpoint prompt template and acquires the evaluation viewpoint prompt. The viewpoint acquisition unit 134 provides the evaluation viewpoint prompt to the generation AI. Then, it is assumed that the viewpoint acquisition unit 134 has obtained, for example, evaluation viewpoints such as "technical feasibility," "user convenience," and "market needs."

[0275] Next, the idea acquisition unit 135 performs the idea acquisition process described using the flowchart in FIG. 9 as follows. That is, the idea acquisition unit 135 acquires an idea generation prompt template from the prompt management table (FIG. 14). The idea acquisition unit 135 also acquires the acquired agent information, including the names "Mr. A, Mr. S, and Mr. X" held by the agent information other than the moderator. The idea acquisition unit 135 also acquires the theme "Solution using AI and cameras." The idea acquisition unit 135 also acquires the acquired idea perspectives "Market entry" and "Technological innovativeness" and evaluation perspectives "Technical feasibility," "User convenience," and "Market needs." Next, the idea acquisition unit 135 assigns "Name," "Theme," "Idea perspective," and "Evaluation perspective" to the respective variables of the idea generation prompt template for each of the three names, thereby acquiring three idea generation prompts. The idea acquisition unit 135 then provides each of the three idea generation prompts to the generation AI and acquires an answer from the generation AI. Next, the idea acquisition unit 135 acquires ideas from each of the three answers. As a result, the idea acquisition unit 135 acquires the ideas "smart assist glasses," "AI-driven traffic monitoring system," and "personalized retail experience system."

[0276] Next, the output unit 14 sequentially transmits the prompts and ideas obtained from the answers to the terminal device 2.

[0277] Next, the terminal device 2 receives and outputs the prompts, answers, etc. sent from the creation support device 1. An example of such output is shown in FIG.

[0278] The current management unit 114 of the creation support device 1 stores the phase identifier "idea generation." In other words, the current phase in the flow of idea creation is "idea generation." As shown in 1601 in FIG. 16, the output unit 14 transmits the answer (idea, etc.) passed from the generation AI in a manner that allows the current "idea generation" identified by the current identifier to be grasped.

[0279] Next, the idea acquiring unit 135 determines that the first person to input an idea exists. Then, the output unit 14 transmits screen information to the terminal device 2 of the user U, prompting the user U to input an idea.

[0280] Next, the terminal device 2 of the user U receives and outputs the screen information. An example of such output is shown in Fig. 17. Then, the user enters an idea such as "customer estimation smart cash register system" into the idea input field 1701 and presses the send button 1702.

[0281] Next, the idea acquisition unit 135 receives an idea such as "customer estimation smart cash register system" from the terminal device 2. Next, the analysis unit 138 analyzes one or more ideas received from the user and determines that they are "normal." As a result, the idea acquisition unit 135 acquires the received idea such as "customer estimation smart cash register system."

[0282] Next, the control unit 132 acquires two transition conditions for moving from the idea generation phase to the next phase from the transition management table. Then, the control unit 132 acquires the number of ideas "4", which is a parameter of the transition conditions. The control unit 132 also acquires "Someone has an idea". Next, the control unit 132 determines that neither of the two transition conditions is satisfied.

[0283] Then, it is assumed that the idea acquisition unit 135 acquires idea generation prompts requested by the moderator from each of A, S, and X, passes each prompt to the generation AI, and acquires answers. Next, it is assumed that the idea acquisition unit 135 acquires the ideas "smart security monitoring system," "AI-driven real-time traffic optimization system," and "AI-driven personal shopping assistant" from each of the three answers, respectively.

[0284] The output unit 14 also transmits to the terminal device 2 the prompts given to the generating AI and the ideas obtained from the answers.

[0285] Next, the terminal device 2 receives and outputs the prompt given to the generating AI. The terminal device 2 also receives and outputs the idea.

[0286] Next, the control unit 132 acquires two transition conditions for moving from the idea generation phase to the next phase from the transition management table (FIG. 14), and determines that the acquired number of ideas, "7," satisfies the transition condition "number of ideas >= x3." In other words, "x3" is a natural number equal to or less than 7.

[0287] Next, the control unit 132 refers to the transition management table and acquires the next phase identifier "idea criticism" that pairs with the transition condition "number of ideas >= x3".

[0288] Then, the control unit 132 updates the current in the current management unit 114 to "idea criticism."

[0289] Next, the idea critique unit 136 performs the idea critique process described using the flowchart in FIG. 11. That is, the idea critique unit 136 acquires an idea critique prompt template from the prompt management table (FIG. 14). The idea critique unit 136 also acquires the names "Mr. A, Mr. S, Mr. X" included in each piece of agent information. The idea critique unit 136 also acquires the theme "Solution using AI and a camera." The idea critique unit 136 also acquires the seven acquired ideas. The idea critique unit 136 assigns "name," "theme," and "idea" to the idea critique prompt template for each idea and for each agent name, thereby acquiring an idea critique prompt. For example, the idea critique unit 136 acquires an idea critique prompt such as "You are Mr. A. Regarding the 'Solution using AI and a camera,' please list any issues or problems with the idea 'Smart Assist Glasses.'" Next, the idea critique unit 136 passes each idea critique prompt to the generation AI. The output unit 14 transmits the prompt given to the generation AI to the terminal device 2. In addition, the idea critique unit 136 acquires an answer from the generation AI. The idea critique unit 136 acquires one or more critiques from the answer from the generation AI. The output unit 14 transmits the acquired one or more critiques to the terminal device 2. The output unit 14 transmits the one or more critiques passed from the generation AI in a manner that allows the current identified by the current identifier to be grasped.

[0290] The terminal device 2 receives and outputs the idea criticism prompt. The terminal device 2 also receives and outputs one or more criticisms for each idea and for each agent name. An example of such output is shown in FIG. 18. 1801 in FIG. 18 is "Mr. A's" criticism of the idea "Smart Assist Glasses." Also, in FIG. 18, it is clearly indicated that the current item is "Idea Criticism" (1802).

[0291] In the idea critique phase, the above steps are repeated for each idea and each agent.

[0292] Next, the control unit 132 acquires the transition condition "number of reviews >= x5" for moving from the idea criticism phase to the next phase from the transition management table (FIG. 13), and determines that the transition condition is satisfied. It is assumed that the control unit 132 acquires the number of acquired reviews, "21," and determines that the condition "number of reviews >= x5" is satisfied.

[0293] Next, the idea selection unit 137 performs the idea selection process described using the flowchart in FIG. 12. That is, the idea selection unit 137 acquires an idea selection prompt template from the prompt management table (FIG. 14). The idea selection unit 137 acquires all of the ideas acquired by the idea acquisition unit 135 and accepted by the idea acceptance unit 122. Here, the set of ideas is seven: "Smart Assist Glasses," "AI-Driven Traffic Monitoring System," "Personalized Retail Experience System," "Customer Estimation Smart Cash Register System," "Smart Security Monitoring System," "AI-Driven Real-Time Traffic Optimization System," and "AI-Driven Personal Shopping Assistant." The idea selection unit 137 also acquires the acquired 21 reviews. Next, the idea selection unit 137 assigns the seven ideas to the variable <Idea Set> of the idea selection prompt template and assigns the acquired 21 reviews to the variable <Previous Reviews> to acquire an idea selection prompt. Next, the idea selection unit 137 passes the idea selection prompt to the generation AI. Then, the idea selection unit 137 acquires an answer from the generation AI. Next, the idea selection unit 137 acquires one or more ideas from the answer. Here, the idea selection unit 137 acquires seven ranked ideas from the answer.

[0294] The output unit 14 transmits the idea selection prompt to the terminal device 2. In addition, the idea output unit 142 transmits information on the seven ranked ideas acquired by the idea selection unit 137 to the terminal device 2.

[0295] The terminal device 2 receives and outputs the idea selection prompt. The terminal device 2 also receives and outputs information on seven ranked ideas. An example of such output is shown in FIG. 19. In FIG. 19, the idea selection prompt is 1901, and the information on the seven ranked ideas is 1902. Also, in FIG. 19, it is clearly indicated that the current is "idea selection" (1903).

[0296] As described above, according to this embodiment, it is possible to support the creation of appropriate ideas using a generation AI.

[0297] Furthermore, according to this embodiment, collaboration between people and the generative AI can help create appropriate ideas.

[0298] Furthermore, according to this embodiment, input by an inappropriate person can be excluded. Furthermore, according to this embodiment, when an input by an inappropriate person is detected, the person can be notified that the input is inappropriate.

[0299] Furthermore, according to this embodiment, the moderator agent can assist in creating more appropriate ideas using the generation AI.

[0300] Furthermore, according to this embodiment, the moderator agent can provide appropriate prompts to the generation AI at appropriate timing.

[0301] Furthermore, according to this embodiment, the flow of discussion for generating appropriate ideas can be presented in an easy-to-understand manner using a generative AI.

[0302] In this embodiment, as described above, the score acquisition unit 139 may acquire scores for one or more ideas acquired by the idea acquisition unit 135 or the idea selection unit 137, and the output unit 14 may output the scores for each of the ideas. For example, if the theme is "new soft drink project ideas," the idea acquisition unit 135 acquires five ideas through the above-described process: "Tropical Coconut (expected to grow due to increasing inbound demand)," "Refreshing Sadistic (demand is expected to increase due to nutritional deficiencies among young people)," "Refreshing Grapefruit (demand is expected to increase due to the aging population)," "Secret Orange (expected to grow due to the increasing popularity of e-sports)," and "Anonymous Grape (office population is at an all-time high, but differentiation from other companies is a challenge)." Next, the idea critique unit 136 acquires one or more critiques for each of the five ideas and for each evaluation criteria through the above-described process. Next, the score acquisition unit 139 ranks the five ideas for each evaluation criteria through the above-described process. Here, it is assumed that the evaluation criteria are "market size," "growth potential," and "barriers to entry." The score acquisition unit 139 then acquires the rankings of the five ideas for each of the three criteria as follows: 1.Market size 1. Refresh Sadistic (The youth nutritional supplement market is the largest at approximately 300 billion yen) 2nd place: Refreshing Grapefruit (The beverage market for seniors is relatively large at approximately 80 billion yen) 3. Anonymous Grape (medium-sized office drinking water market at 110 billion yen) 4. Tropical Coconut (The specific market size for resort drinks is unknown) 5. Secret Orange (The esports market is relatively small at ¥143 billion) 2. Growth potential 1st place: Tropical Coconut (expected to grow due to rising inbound demand) 2. Refresh Sadistic (Demand is expected to increase due to nutritional deficiencies among young people) 3. Refreshing Grapefruit (Demand is expected to increase due to the aging population) 4. Secret Orange (Expected to grow due to the growing popularity of esports) 5. Anonymous Grape (Office population is at an all-time high, but differentiation from other companies is an issue) 3. Barriers to entry? 1. Anonymous Grape (Many similar plain plastic bottle products are available) 2. Refresh Sadistic (Difficult to differentiate as many manufacturers have already entered the market) 3. Tropical coconut (stable supply is an issue due to seasonal and regional fluctuations in demand) 4. Refreshing Grapefruit (a market in which major manufacturers have already entered) 5. Secret Orange (Although similar products exist, there is room for new entrants)

[0303] Next, the score acquisition unit 139 calculates the score for each evaluation criteria for each idea using the above-mentioned formula "Score for each evaluation criteria = (100 * (number of ideas - ranking)) / (number of evaluation criteria x (number of ideas - 1))". Next, the score acquisition unit 139 calculates the sum of the scores for each of the three evaluation criteria for each of the five scores and ranks them. Then, the output unit outputs the ranking and score (total score) of each of the five ideas as shown in the "overall evaluation" below. comprehensive evaluation 1st place: Refresh Sadistic 33.3 + 25.0 + 25.0 = 83.3 2nd place: Tropical Coconut 8.3 + 33.3 + 16.6 = 58.2 3rd place: Refreshing Grapefruit 25.0 + 16.6 + 8.3 = 49.9 4th place: Anonymous Grape 16.6 + 0.0 + 33.3 = 49.9 5. Secret Orange 0.0 + 8.3 + 0.0 = 8.3

[0304] The scores for each evaluation criteria were: 1st place: 33.3 points, 2nd place: 25.0 points, 3rd place: 16.6 points, 4th place: 8.3 points, and 5th place: 0.0 points.

[0305] The processing in this embodiment may also be realized by software. This software may be distributed by software download or the like. This software may also be recorded on a recording medium such as a CD-ROM and distributed. This also applies to other embodiments in this specification. The software that realizes the creation support device 1 in this embodiment is the following program. That is, this program causes a computer that can access a prompt management unit that stores setting prompt templates for configuring prompts for character setting of one or more agents, and phase prompt templates for configuring prompts to be given to the generation AI in each of two or more phases including an idea generation phase and an idea selection phase, and an attribute value storage unit that stores one or more idea attribute values ​​that are attribute values ​​related to ideas to be created, The program functions as a character setting unit that acquires one or more setting prompts using the setting prompt template and gives each of the one or more setting prompts to the generation AI; an idea acquisition unit that acquires a prompt using the phase prompt template corresponding to the idea generation phase and the one or more idea attribute values ​​and gives the prompt to each of the one or more agents of the generation AI to acquire two or more ideas; an idea selection unit that acquires a prompt using the phase prompt template corresponding to the idea selection phase and the two or more ideas acquired by the idea acquisition unit and gives the prompt to the generation AI to acquire one or more ideas selected based on the two or more ideas acquired by the idea acquisition unit; and an idea output unit that outputs the one or more ideas acquired by the idea selection unit.

[0306] (Embodiment 2) In this embodiment, we will describe a discussion support device in which two or more parties, including two or more generation AI personas including a moderator, hold a discussion on a predetermined topic, and when a predetermined condition is met, the discussion support device passes a prompt from the moderator persona to the generation AI and outputs utterance information based on the response from the generation AI. Note that the predetermined condition is, for example, that the utterance information of all parties except the moderator has been output.

[0307] In addition, in this embodiment, a discussion support device that can automatically set the required AI persona will be described.

[0308] The conceptual diagram of information system B in this embodiment is the same as that in Fig. 1 except for the reference numerals. In information system B, it is preferable that discussion support device 4 has all the components of creation support device 1. It can be said that support for idea creation performed by creation support device 1 is support for discussion on a predetermined theme performed by discussion support device 4.

[0309] The conference support device 4 is, for example, a cloud server or an ASP server, but the type does not matter. The conference support device 4 may also be a terminal. If the conference support device 4 is a terminal, the information system B does not require the terminal device 2. If the conference support device 4 is a terminal, it can be considered that the conference support device 4 also serves as the terminal device 2. If the conference support device 4 is a terminal, the conference support device 4 is, for example, a so-called personal computer, smartphone, or tablet terminal, but the type does not matter.

[0310] FIG. 20 is a block diagram of the consultation support device 4 constituting the information system B in this embodiment.

[0311] The consultation support device 4 includes a storage unit 41, a reception unit 42, a processing unit 43, and an output unit 44. The storage unit 41 includes a transition management unit 111, a prompt management unit 412, an attribute value storage unit 113, and a current management unit 114. The reception unit 42 includes an attribute value reception unit 421 and an idea reception unit 122. The processing unit 43 includes an agent determination unit 130, a character setting unit 131, a control unit 132, a moderator judgment unit 133, a viewpoint acquisition unit 134, an idea acquisition unit 135, an idea criticism unit 136, an idea selection unit 137, an analysis unit 138, a score acquisition unit 139, and a comment acquisition unit 431. The output unit 44 includes a response output unit 141, an idea output unit 142, a score output unit 143, and an error output unit 144.

[0312] Various types of information are stored in the storage unit 41. The various types of information include, for example, transition information and prompt templates.

[0313] Various prompt templates are stored in the prompt management unit 412. The various prompt templates include, for example, a setting prompt template, one or more agent determination prompt templates, and one or more moderator prompt templates. Preferably, each of the two or more moderator prompt templates corresponds to a phase identifier of a different phase.

[0314] The setting prompt template here is information for configuring a prompt for character setting of two or more agents, including the moderator's agent. An example of a setting prompt template for setting the moderator's agent is, "Mr. X is the moderator for holding a discussion on <topic>."

[0315] A moderator prompt template is a template for constructing a prompt to be given to the generation AI to proceed with a discussion on a topic. A moderator prompt template is a template for the moderator. Examples of moderator prompt templates are, "I'm X, the moderator. Participants, please state your opinions on <topic>.", "I'm X, the moderator. Participants, after listening to the opinions of the others, do you have any other opinions?", and "I'm X, the moderator. Participants, please summarize your opinions from your respective perspectives."

[0316] Note that "I'm X, the moderator. Participants, please give your opinions on <topic>." is a moderator prompt template that corresponds to the first phase. "I'm X, the moderator. Participants, after listening to the opinions of the others, do you have any other opinions?" is a moderator prompt template that corresponds to the second phase. Also, "I'm X, the moderator. Participants, please summarize your opinions from your respective perspectives." is a moderator prompt template that corresponds to the third phase.

[0317] The reception unit 42 receives various instructions and information, such as a theme and a start instruction.

[0318] Here, acceptance typically refers to the reception of information transmitted via a wired or wireless communication line, but may also be a concept that includes the reception of information input from an input device such as a keyboard, mouse, or touch panel, or the reception of information read from a recording medium such as an optical disk, magnetic disk, or semiconductor memory.

[0319] The attribute value receiving unit 421 receives a theme. The theme is a topic for discussion. The theme may be the theme for which ideas are generated in the first embodiment, or may be a theme unrelated to idea creation. Examples of themes unrelated to idea creation are "trends among young people today" and "best memories from school days."

[0320] The idea receiving unit 122 receives one or more pieces of comment information from a person. The comment information may be the idea in the first embodiment.

[0321] The processing unit 43 performs various types of processing, such as processing performed by the agent determination unit 130, the character setting unit 131, the control unit 132, the moderator judgment unit 133, the viewpoint acquisition unit 134, the idea acquisition unit 135, the idea criticism unit 136, the idea selection unit 137, the analysis unit 138, the score acquisition unit 139, or the comment acquisition unit 431.

[0322] The moderator determination unit 133 determines whether the prompt given to the generation AI or the response from the generation AI satisfies the moderator's statement condition. The moderator's statement condition may be any transition condition.

[0323] Moderator speech conditions are conditions for the moderator agent to decide to speak. Moderator speech conditions are conditions related to the prompt given to the generation AI or conditions related to the response from the generation AI. Moderator speech conditions are, for example, "Speech information from all two or more parties has been obtained," "Speech information from all two or more parties including people has been obtained," "The number of speech information related to the topic has exceeded or exceeded a threshold," and "A specified amount of time has passed without any input from the people involved."

[0324] The utterance acquisition unit 431 acquires a prompt for obtaining utterance information related to the theme, provides the prompt to the generation AI, and acquires utterance information from one or more of the two or more agents. The utterance acquisition unit 431, for example, creates a prompt using a moderator prompt template and a theme, provides the prompt to the generation AI, and acquires utterance information from one or more of the two or more agents. The utterance information may be an answer from the generation AI, or may be information acquired from an answer from the generation AI. The utterance information is information indicating an agent's utterance related to the theme. The utterance information is, for example, information about an opinion related to the theme. The utterance information is usually a string of characters. The utterance information may also be the above-mentioned idea.

[0325] When the moderator determination unit 133 determines that the moderator utterance conditions are met, the utterance acquisition unit 431 preferably acquires a prompt based on the moderator prompt template, provides the prompt to the generation AI, and acquires one or more pieces of utterance information. A prompt based on the moderator prompt template is usually a prompt for one or more agents other than the moderator.

[0326] The output unit 44 outputs one or more pieces of utterance information acquired by the utterance acquisition unit 431. Here, output usually means transmission to the terminal device 2, but may also be a concept that includes display on a display, projection using a projector, printing on a printer, sound output, storage on a recording medium, delivery of processing results to another processing device, another program, etc.

[0327] The storage unit 41 is preferably a non-volatile recording medium, but can also be realized as a volatile recording medium.

[0328] There is no restriction on the process by which information is stored in the storage unit 41. For example, information may be stored in the storage unit 41 via a recording medium, information transmitted via a communication line or the like may be stored in the storage unit 41, or information input via an input device may be stored in the storage unit 41.

[0329] The reception unit 42 and the attribute value reception unit 421 are usually realized by wireless or wired communication means, but may also be realized by a means for receiving broadcasts, a device driver for an input means such as a touch panel or keyboard, or control software for a menu screen.

[0330] The processing unit 43 and the statement acquisition unit 431 can usually be realized by a processor, memory, etc. The processing procedures of the processing unit 43, etc. are usually realized by software, and the software is recorded on a recording medium such as a ROM. However, they may also be realized by hardware (dedicated circuit). The processor may be a CPU, MPU, GPU, etc., and the type does not matter.

[0331] The output unit 44 is usually realized by a wireless or wired communication means, but may also be realized by driver software for an output device such as a display or speaker, or by a combination of driver software for an output device and the output device.

[0332] Next, an example of the operation of the conference support device 4 will be described with reference to the flowchart of FIG.

[0333] (Step S2101) The attribute value receiving unit 421 determines whether or not a start instruction has been received. If a start instruction has been received, the process proceeds to step S2102, and if not, the process returns to step S2101. The start instruction usually includes a theme.

[0334] (Step S2102) The processing unit 43 performs pre-processing. An example of the pre-processing will be described with reference to the flowchart in FIG.

[0335] (Step S2103) The processing unit 43 assigns 1 to the counter i.

[0336] (Step S2104) The processing unit 43 determines whether or not the i-th phase exists. If the i-th phase exists, the process proceeds to step S2105; if not, the process returns to step S2101. The processing unit 43 determines, for example, whether or not a phase identifier that identifies the i-th phase exists in the prompt management unit 412. Each phase identifier in the prompt management unit 412 is associated with, for example, one or more moderator prompt templates.

[0337] (Step S2105) The utterance acquisition unit 431 acquires a moderator prompt template corresponding to the i-th phase from the prompt management unit 412. The utterance acquisition unit 431 acquires a moderator prompt using the moderator prompt template. Note that the utterance acquisition unit 431 acquires the moderator prompt, for example, by substituting the theme of the start instruction for the variable <theme> of the moderator prompt template. Note that the moderator prompt template and the moderator prompt may be the same.

[0338] (Step S2106) The statement acquisition unit 431 passes the moderator prompt acquired in step S2105 to the generation AI.

[0339] (Step S2107) The comment acquisition unit 431 acquires comment information based on the response from the generation AI. An example of such comment acquisition processing will be described with reference to the flowchart in FIG.

[0340] (Step S2108) The moderator determination unit 133 determines whether the moderator's utterance conditions are met. If the moderator's utterance conditions are met, proceed to step S2109; if the moderator's utterance conditions are not met, return to step S2107. When returning to step S2107, the utterance acquisition unit 431 may pass to the generation AI the same prompt as the prompt previously passed to the generation AI. In other words, if the moderator's utterance conditions are not met, return to step S2106.

[0341] (Step S2109) The processing unit 43 increments the counter i by 1. The process returns to step S2104.

[0342] In the flowchart of FIG. 21, the process ends when the power is turned off or an interrupt occurs to end the process.

[0343] Next, an example of the pre-processing in step S2102 will be described with reference to the flowchart in Fig. 22. Note that each step in the flowchart in Fig. 22 is configured with each step in Fig. 5, and therefore description thereof will be omitted.

[0344] Next, an example of the comment acquisition process in step S2107 will be described with reference to the flowchart in FIG.

[0345] (Step S2301) The message acquisition unit 431 assigns 1 to a counter i.

[0346] (Step S2302) The statement acquisition unit 431 determines whether or not the i-th agent exists. If the i-th agent exists, the process proceeds to step S2303; if not, the process proceeds to step S2307. The i-th agent is an agent excluding the moderator.

[0347] (Step S2303) The statement acquisition unit 431 determines whether or not the reply of the i-th agent has been acquired from the generation AI. If the reply of the i-th agent has been acquired, the process proceeds to step S2304; if not, the process returns to step S2303.

[0348] (Step S2304) The utterance acquiring unit 431 acquires utterance information from the reply of the i-th agent. Note that the utterance information is usually a character string included in the reply. However, the reply and the utterance information may be the same.

[0349] (Step S2305) The output unit 44 outputs the utterance information acquired in step S2304 in a manner that clearly indicates that the utterance is from the i-th agent.

[0350] (Step S2306) The message acquiring unit 431 increments the counter i by 1. The process returns to step S2302.

[0351] (Step S2307) The comment acquisition unit 431 determines whether or not to accept input from a person. If input from a person is accepted, the process proceeds to step S2308, and if input is not accepted, the process returns to the upper level process. Note that whether or not to accept input from a person is usually determined in advance.

[0352] (Step S2308) The message acquisition unit 431 assigns 1 to the counter i.

[0353] (Step S2309) The comment acquisition unit 431 determines whether or not there is an i-th person who will input comment information. If there is an i-th person, the process proceeds to step S2310; if there is not, the process returns to the upper level process.

[0354] (Step S2310) The comment acquisition unit 431 determines whether or not comment information has been received from the i-th person. If comment information has been received, the process proceeds to step S2311, and if not, the process returns to step S2310. Note that if a timeout occurs here, the process may proceed to S2313. Also, the reception from a person may be, for example, the reception of comment information from the terminal device 2.

[0355] (Step S2311) The comment acquisition unit 431 acquires comment information from the i-th person.

[0356] (Step S2312) The output unit 44 outputs the utterance information acquired in step S2311.

[0357] (Step S2313) The message acquiring unit 431 increments the counter i by 1. The process returns to step S2302.

[0358] As described above, according to this embodiment, by setting a moderator agent, the progress of a discussion on a predetermined theme can be made smooth using the generating AI.

[0359] Furthermore, according to this embodiment, when the moderator's statement conditions are met, a prompt is given to the generation AI and one or more pieces of statement information are obtained, which allows the progress of discussion on a predetermined topic to proceed smoothly.

[0360] Furthermore, according to this embodiment, it is possible to automatically set an appropriate agent for advancing a discussion on a predetermined topic using the generation AI.

[0361] The software for realizing the discussion support device 4 is the following program: That is, this program includes a computer that can access a prompt management unit that stores setting prompt templates for configuring prompts for character settings of two or more agents including a moderator's agent, and one or more moderator prompt templates that are templates for configuring prompts to be given to the generation AI for the progress of the discussion on the theme and are templates for the moderator, an attribute value receiving unit that receives a theme, a character setting unit that obtains two or more setting prompts using the setting prompt templates and gives each of the one or more setting prompts to the generation AI, and a character setting unit that obtains prompts using the moderator prompt template and the theme, gives the prompts to the generation AI, and sets the two or more setting prompts to the generation AI. The program causes the computer to function as a statement acquisition unit that acquires statement information from one or more of the above agents, an output unit that outputs the one or more pieces of statement information acquired by the statement acquisition unit, and a moderator judgment unit that determines whether the moderator statement conditions, which are conditions for the moderator's agent to speak and are conditions related to the prompt given to the generation AI or conditions related to the answer from the generation AI, are met, and when the moderator judgment unit determines that the moderator statement conditions are met, the statement acquisition unit acquires a prompt based on the moderator prompt template, gives the prompt to the generation AI, and acquires one or more pieces of statement information.

[0362] FIG. 24 is a block diagram of a computer system 300 that executes the programs described in this specification to realize the creation support device 1, the discussion support device 4, and the like according to the various embodiments described above.

[0363] In FIG. 24, a computer system 300 includes a computer 301 including a CD-ROM drive, a keyboard 302, a mouse 303, and a monitor 304.

[0364] 24, computer 301 includes, in addition to CD-ROM drive 3012, MPU 3013, bus 3014 connected to CD-ROM drive 3012 etc., ROM 3015 for storing programs such as a boot-up program, RAM 3016 connected to MPU 3013 for temporarily storing instructions of application programs and providing temporary storage space, and hard disk 3017 for storing application programs, system programs, and data. Although not shown here, computer 301 may further include a network card for providing connection to a LAN.

[0365] A program that causes the computer system 300 to execute the functions of the creation support device 1 of the above-described embodiment may be stored on a CD-ROM 3101, inserted into the CD-ROM drive 3012, and then transferred to the hard disk 3017. Alternatively, the program may be transmitted to the computer 301 via a network (not shown) and stored on the hard disk 3017. The program is loaded into the RAM 3016 when executed. The program may also be loaded directly from the CD-ROM 3101 or the network.

[0366] The program does not necessarily include an operating system (OS) or a third-party program that causes the computer 301 to execute the functions of the creation support device 1 of the above-described embodiment. The program only needs to include instructions that call appropriate functions (modules) in a controlled manner to achieve the desired results. How the computer system 300 operates is well known, and a detailed description thereof will be omitted.

[0367] In addition, in the above program, the steps of transmitting information and receiving information do not include processing performed by hardware, such as processing performed by a modem or interface card in the transmission step (processing that can only be performed by hardware).

[0368] The computer that executes the program may be a single computer or a plurality of computers, that is, it may perform centralized processing or distributed processing.

[0369] Furthermore, in each of the above embodiments, it goes without saying that two or more communication means present in one device may be physically realized by one medium.

[0370] Furthermore, in each of the above embodiments, each process may be realized by centralized processing in a single device, or may be realized by distributed processing in a plurality of devices.

[0371] The present invention is not limited to the above-described embodiment, and various modifications are possible, and it goes without saying that these modifications are also included within the scope of the present invention. [Industrial Applicability]

[0372] As described above, the creation support device 1 according to the present invention has the effect of being able to support the creation of appropriate ideas using a generation AI, and is useful as a server or the like for idea creation. [Explanation of symbols]

[0373] A. Information Systems 1 Creation support device 2. Terminal Device 3 Generation AI device 4 Discussion support device 11, 41 Storage area 12, 42 Reception 13, 43 Processing section 14, 44 Output section 21 Terminal storage section 22 Terminal Reception 23 Terminal processing section 24 Terminal transmitter 25 Terminal receiving section 26 Terminal Output Unit 111 Transition Management Department 112 Prompt Management Department 113 Attribute Value Storage 114 Current Management Department 121, 421 Attribute value reception section 122 Idea Reception Department 130 Agent Determination Unit 131 Person setting section 132 Control Unit 133 Chairperson Judgment Department 134 Perspective Acquisition Department 135 Idea Acquisition Department 136 Idea Criticism Department 137 Idea Selection Department 138 Analysis Department 139 Score Acquisition Section 141 Answer output section 142 Idea Output Unit 143 Score output section 144 Error output section 431 Speech Acquisition Department

Claims

1. a computer that can access a prompt management unit that stores setting prompt templates for configuring prompts for character settings of two or more agents including a moderator's agent, and one or more moderator prompt templates that are templates for configuring prompts to be given to the generation AI for the progress of a discussion on the theme and are templates for the moderator; an attribute value receiving unit that receives a theme; a person setting unit that obtains two or more setting prompts using the setting prompt template and provides each of the one or more setting prompts to the generation AI; a utterance acquisition unit that acquires a prompt using the moderator prompt template and the theme, provides the prompt to the generation AI, and acquires utterance information from one or more of the two or more agents; A program for causing the computer to function as an output unit that outputs one or more pieces of utterance information acquired by the utterance acquisition unit.

2. The computer further functioning as an agent determination unit that determines one or more agents according to the theme; The person setting unit 3. The program according to claim 1, for causing the computer to function as a program for obtaining setting prompts for the one or more agents determined by the agent determination unit and providing the setting prompts to the generation AI.

3. A discussion support device in which two or more parties, including two or more agents of a generation AI, hold a discussion on a predetermined theme, a prompt management unit in which setting prompt templates for configuring prompts for character settings of two or more agents including a moderator's agent, and one or more moderator prompt templates, which are templates for configuring prompts to be given to the generation AI for the progress of the discussion on the theme and are templates for the moderator, are stored; an attribute value receiving unit that receives a theme; a person setting unit that obtains two or more setting prompts using the setting prompt template and provides each of the one or more setting prompts to the generation AI; a utterance acquisition unit that acquires a prompt using the moderator prompt template and the theme, provides the prompt to the generation AI, and acquires utterance information from one or more of the two or more agents; an output unit that outputs one or more pieces of utterance information acquired by the utterance acquisition unit.

4. A discussion support method realized by a prompt management unit storing setting prompt templates for configuring prompts for character settings of two or more agents including a moderator's agent, and one or more moderator prompt templates, which are templates for configuring prompts to be given to the generation AI for the progress of a discussion on the theme and are templates for the moderator, an attribute value receiving unit, a character setting unit, a utterance acquisition unit, an output unit, and a moderator determination unit, an attribute value receiving step in which the attribute value receiving unit receives a theme; a person setting step in which the person setting unit obtains two or more setting prompts using the setting prompt template and provides each of the one or more setting prompts to the generation AI; a utterance acquisition step in which the utterance acquisition unit acquires a prompt using the moderator prompt template and the theme, provides the prompt to the generation AI, and acquires utterance information from one or more agents among the two or more agents; and an output step in which the output unit outputs one or more pieces of utterance information acquired by the utterance acquisition unit.

Images