Information processing apparatus, information processing method, program

Abstract

An object of the present invention is to provide a technique capable of evaluating behavior-variable content. 【Means for Solving the Problem】 The information processing apparatus according to the present invention includes: a first agent processing unit that acquires a plurality of reactions obtained from each of a plurality of first agents having different behavior characteristics as a result of causing the same behavior-variable content to act on each of the plurality of first agents; a second agent processing unit that causes a second agent having a predetermined attribute to evaluate the behavior-variable content based on the plurality of reactions obtained from each of the plurality of first agents; and an output unit that outputs an evaluation result of the behavior-variable content.

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] In recent years, efforts to apply an approach based on the theory of behavioral science, which scientifically studies human behavior, to service development have spread in various fields such as public policy, medicine, retail, and education. In addition, systems for technically realizing the support for behavior modification and habituation of target persons in such various fields have also been studied (for example, Patent Document 1).

[0003] In the system described in Patent Document 1, behavioral data including various data measured for the behaviors of a plurality of target persons is analyzed, and based on the results of the analysis of the behavioral data, a stage index that is an index as a criterion for a plurality of stages that gradually lead to the behavior as the goal of habituation, and each of the plurality of stages according to the stage index are defined, for each pair of adjacent stages, identifying the gap between the two stages that make up the pair, for each stage pair, at least one of the reason for the identified gap and the measure for causing a behavior change for a target person belonging to the lower stage to transition to the higher stage from the relationship information in which the relationship between the gap and the reason / measure is defined, and executing processing related to the reason / measure identified for each stage pair is described.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, the system described in Patent Document 1 does not evaluate the effect of behavior modification content, which is a measure for causing behavior modification. Further, even if the effect of behavior modification content is to be evaluated, it is necessary to analyze the behavior data obtained by measuring the behavior when the behavior modification content is actually applied to a plurality of subjects, and it was difficult to evaluate the behavior modification content before actually applying it to the subjects.

[0006] Therefore, an object of the present invention is to provide a technology capable of evaluating behavior modification content.

Means for Solving the Problems

[0007] An information processing apparatus according to an aspect of the present invention includes a first agent processing unit that acquires a plurality of reactions obtained from each of a plurality of first agents having different behavior characteristics as a result of causing the same behavior modification content to act on each of the plurality of first agents, a second agent processing unit that causes a second agent having a predetermined attribute to evaluate the behavior modification content based on the plurality of reactions obtained from each of the plurality of first agents, and an output unit that outputs an evaluation result of the behavior modification content.

[0008] An information processing method according to an aspect of the present invention includes an acquisition step of acquiring a plurality of reactions obtained from each of a plurality of first agents having different behavior characteristics as a result of causing the same behavior modification content to act on each of the plurality of first agents, an evaluation step of causing a second agent having a predetermined attribute to evaluate the behavior modification content based on the plurality of reactions obtained from each of the plurality of first agents, and an output step of outputting an evaluation result of the behavior modification content.

[0009] A program according to an aspect of the present invention causes a computer to execute an acquisition step of acquiring a plurality of reactions obtained from each of a plurality of first agents as a result of causing the same behavior modification content to act on each of the plurality of first agents having different behavior characteristics, an evaluation step of causing a second agent having a predetermined attribute to evaluate the behavior modification content based on the plurality of reactions obtained from each of the plurality of first agents, and an output step of outputting an evaluation result of the behavior modification content.

[0010] According to these aspects, for example, by having another agent evaluate the results of causing behavior modification content to act on a plurality of agents having different behavior characteristics, it is possible to evaluate the behavior modification content without using a group of real-world humans.

[0011] In the above aspect, each of the plurality of first agents may be an agent that reflects, without overlapping each other, a plurality of independent concepts included in a predetermined classification method related to a behavior modification method.

[0012] According to this aspect, since the behavior characteristics of each agent group can efficiently reflect the behavior characteristics of a group of real-world humans without overlap, it becomes possible to evaluate behavior modification content with a smaller number of agents.

[0013] In the above aspect, each of the plurality of first agents may be an agent that comprehensively reflects a plurality of independent concepts included in a predetermined classification method related to a behavior modification method.

[0014] According to this aspect, since it is possible to comprehensively reflect the behavior characteristics of each person in a group of real-world humans as a whole agent group, it becomes possible to evaluate behavior modification content more widely and objectively.

[0015] In the above aspect, it may further include a reception unit that receives predetermined specified content from a user, and the second agent may be an agent that reflects the predetermined specified content received from the user.

[0016] According to this aspect, by giving the second agent a role according to the specified content from the user and having it evaluated, it is possible to enhance the acceptance of the evaluation result.

Effect of the Invention

[0017] According to the present invention, it is possible to provide a technology capable of evaluating behavior modification content.

Brief Description of the Drawings

[0018]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Embodiments for Carrying Out the Invention

[0019] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. The same elements are denoted by the same reference numerals, and redundant descriptions are omitted.

[0020] In this embodiment, an agent is a system or software program for autonomously performing a specific task. The agent may be, for example, an agent that utilizes generative AI such as a large language model (LLM) or a large action model (LAM) (hereinafter referred to as an "AI agent"). The AI agent can interact with the environment, determine the next action to be executed based on the collected data, etc., and act autonomously.

[0021] Also, each agent can be given specific behavioral characteristics. For example, if the AI agent is set with behavioral characteristics such as being proactive and highly curious, the AI agent will tend to act actively and challenge new things based on the set behavioral characteristics. <System Overview> FIG. 1 is a diagram showing an example of the system configuration of the information processing system in this embodiment. In the example shown in FIG. 1, the information processing system 1 includes an information processing device 10, an information processing terminal 20, and a network N.

[0022] The information processing device 10 and the information processing terminal 20 can communicate with each other via the network N. In FIG. 1, one information processing device 10 and two information processing terminals 20 are illustrated, but the numbers of the information processing devices 10 and information processing terminals 20 included in the information processing system 1 are not limited to this, and any number of devices can be included respectively.

[0023] The information processing apparatus 10 is configured by, for example, a server apparatus, a cloud computing form, an ASP (Application Service Provider), a client-server model, etc., but is not limited thereto. The information processing terminal 20 is an information processing terminal device used by a user, and is, for example, a mobile phone terminal (including a smartphone), a tablet, or a personal computer, but is not limited thereto.

[0024] The network N is a communication network for communication between the information processing apparatus 10 and the information processing terminal 20. For example, the network N may be any of the Internet, an intranet, a LAN, a mobile communication network, a dedicated line, a packet communication network, a telephone line, a corporate internal network, other communication lines, combinations thereof, etc. Further, the network N may be wired or wireless.

[0025] <Hardware Configuration> FIG. 2 is a diagram showing an example of the hardware configuration of the information processing apparatus 10 and the information processing terminal 20 in the present embodiment. The information processing apparatus 10 includes a processor 11 such as a CPU (Central Processing Unit) and a GPU (Graphical Processing Unit), a storage device 12 such as a memory, an HDD (Hard Disk Drive), and an SSD (Solid State Drive), a communication IF (Interface) 13 that performs wired or wireless communication, an input device 14 that receives an input operation, and an output device 15 that outputs information. The input device 14 is, for example, a keyboard, a touch panel, a mouse, a microphone, etc. The output device 15 is, for example, a display, a touch panel, a speaker, etc. The information processing terminal 20 also has a similar hardware configuration.

[0026] <Functional Block Configuration> (Information Processing Apparatus 10) FIG. 3 is a diagram showing an example of the functional block configuration of the information processing apparatus 10 in the present embodiment. The information processing apparatus 10 includes a storage unit 100 and a control unit 110. The storage unit 100 can be realized using a storage device 12 included in the information processing apparatus 10.

[0027] Further, the control unit 110 can be realized by a processor 11 of the information processing apparatus 10 executing a program stored in the storage device 12. Also, the program can be stored in a storage medium. The storage medium storing the program may be a computer-readable non-transitory storage medium (Non-transitory computer readable medium). The non-transitory storage medium is not particularly limited, and for example, it may be a storage medium such as a USB memory or a CD-ROM.

[0028] The information processing apparatus 10 has a function of performing various communications with the information processing terminal 20 using the communication IF 13. The information processing apparatus 10 is provided with a communication interface compliant with various communication standards, and transmits and receives various information to and from external devices including each information processing terminal 20. The information processing apparatus 10 receives data transmitted from the information processing terminal 20 via the network N.

[0029] The storage unit 100 stores data necessary for the information processing apparatus 10 to execute the determination process. The data includes behavior characteristic data 40 regarding human behavior characteristics, attribute data 41 representing the attributes of the agent, and behavior modification content 52 for promoting the behavior modification of the target person, etc.

[0030] The control unit 110 provides various functions necessary for executing the determination process. The control unit 110 includes an agent generation unit 111, a first agent processing unit 112, a second agent processing unit 113, a reception unit 114, and an output unit 115, which will be described later.

[0031] Note that the agent generation unit 111, the first agent processing unit 112, and the second agent processing unit 113 may be constructed on the information processing device 10, or may be constructed to operate on a generation AI system (not shown) that is communicable via the network N and provided by an external system.

[0032] For example, an environment in which the agent generation unit 111, the first agent processing unit 112, and the second agent processing unit 113 operate may be constructed on a generation AI service such as ChatGPT (registered trademark), and the processing of the agent generation unit 111, the first agent processing unit 112, and the second agent processing unit 113 may be executed.

[0033] The agent generation unit 111 has a function of generating a plurality of first agents 51 having different behavioral characteristics based on the behavioral characteristic data 40. The first agents 51 are independent of each other. The agent generation unit 111 can set the behavioral characteristics of each first agent 51 based on a specific concept of the academic classification method of the behavior variation method. The agent generation unit 111 may compile an agent group in which each of the plurality of first agents 51 mimics the concepts of the academic classification method of the behavior variation method without omission and duplication. By compiling an agent group in which the plurality of first agents 51 collectively mimic the concepts of the academic classification method of the behavior variation method without omission, that is, comprehensively, the behavioral characteristics of the entire population of real-world humans can be reflected comprehensively as the entire agent group. This makes it possible to evaluate behavior variation content more widely and objectively. In addition, by compiling an agent group in which each of the plurality of first agents 51 mimics the concepts of the academic classification method of the behavior variation method without duplication, each agent can efficiently reflect the behavioral characteristics of the population of real-world humans without duplication. This makes it possible to evaluate behavior variation content with a smaller number of agents. Note that the plurality of generated first agents 51 are also referred to as the crowd AI 50.

[0034] Although various action characteristics can be applied, in this embodiment, it is preferable to set the action characteristics of each agent based on the academic classification method of the action transformation method. Examples of the academic classification method of the action transformation method include the following. · Integrated Behavioral Model (IBM) · Health Belief Model (HBM) · Big Five personality traits

[0035] For example, when defining the action characteristic data 40 based on the Big Five personality traits, as the action characteristic data 40, the strength of the tendency regarding each factor (extroversion, agreeableness, conscientiousness, neuroticism, openness to experience) of the Big Five personality traits is quantified. It may be considered that the larger the numerical value, the stronger the tendency regarding the factor.

[0036] In addition, the agent generation unit 111 has a function of generating one or a plurality of second agents 61 having a predetermined attribute based on the attribute data 41. The attributes include, for example, gender, age, educational and work history, annual income, lifestyle, family composition, hobbies, values, and the like. The agent generation unit 111 generates a second agent imitating a specific person, such as a politician or a 30-year-old man living in the city, by setting a predetermined attribute. Note that one or a plurality of second agents 61 having a predetermined attribute are also collectively referred to as the AI 60.

[0037] The first agent processing unit 112 acts on each of the plurality of first agents 51 with behavior modification content 52, and as a result of the action, acquires a plurality of reactions obtained from each of them. For example, the first agent processing unit 112 constructs a simulation space imitating human society on the AI system for generation, and acts on the crowd AI 50 generated by the agent generation unit 111 with the same behavior modification content 52 for each. An example of the behavior modification content 52 is a nudge message. Each first agent 51 constituting the crowd AI 50 makes a reaction such as answering in the AI system for generation a judgment such as causing or not causing an action with respect to the nudge message based on the behavior characteristics set for each.

[0038] The first agent processing unit 112 may measure, for each reaction of the crowd AI 50 and each first agent, the time until the reaction, the amount of information consumed, and the like.

[0039] The second agent processing unit 113 acquires reaction information 53 from the first agent processing unit 112. The second agent processing unit 113 evaluates the behavior modification content 52 by having one or a plurality of second agents 61 evaluate the reaction information 53.

[0040] In the present embodiment, the behavior modification content 52 is content intended to cause a user to execute a predetermined target behavior. That is, the behavior modification content 52 is, for example, content having an effect of modifying the behavior of the user.

[0041] As an example of the behavior modification content 52, for example, there is a nudge message. A nudge message is a single or a plurality of messages that influence people's decision-making with a small trigger based on behavioral science without greatly changing economic incentives or forcing actions with penalties or rules, and promote behavior modification. Note that the behavior modification content 52 is not limited to a text format including character strings and symbols, and may be predetermined information including voice, video, or images.

[0042] The second agent 61 outputs, as an evaluation, for example, the ratio of the first agents that have taken the target action, the reasons why the first agents have not taken the target action and their ratios, etc.

[0043] The reception unit 114 receives instructions and inputs from the user. For example, the reception unit 114 can receive instructions and inputs from the user regarding the action characteristics and attributes of the agent to be generated, and the action modification content 52.

[0044] The output unit 115 outputs the evaluation result 70 of the action modification content 52. The output unit 115 can, for example, transmit the evaluation result 70 of the action modification content 52 to the information processing terminal 20.

[0045] (Information processing terminal 20) FIG. 4 is a diagram showing an example of the functional block configuration of the information processing terminal 20 according to the present embodiment. The information processing terminal 20 includes a storage unit 200, a communication unit 201, a UI (User Interface) unit 202, and a control unit 203. The storage unit 200 can be realized using a storage device provided in the information processing terminal 20. Further, the communication unit 201, the UI unit 202, and the control unit 203 can be realized by a processor of the information processing terminal 20 executing a program stored in the storage device. Further, the program can be stored in a storage medium. The storage medium storing the program may be a computer-readable non-transitory storage medium. The non-transitory storage medium is not particularly limited, but may be, for example, a storage medium such as a USB memory or a CD-ROM.

[0046] The storage unit 200 stores various programs and data necessary for the control unit 203 to execute this information processing.

[0047] The communication unit 201 has a function of performing various communications with the information processing apparatus 10 using a communication IF. The communication unit 201 is provided with a communication interface compliant with various communication standards, and transmits and receives various information to and from external devices including each information processing apparatus 10.

[0048] <Agent Processing> FIG. 5 is an explanatory diagram for explaining a series of processes for evaluating the behavior modification content 52 in the present embodiment. In the present embodiment, the behavior modification content 52 is evaluated by collecting the results of reacting the same behavior modification content 52 with each first agent 51 and having each second agent 61 evaluate it.

[0049] First, the agent generation unit 111 composes the first agent 51 and the second agent 61. The agent generation unit 111 generates a plurality of first agents 51 having different behavior characteristics based on the behavior characteristic data 40. The behavior characteristic data 40 is data related to human behavior characteristics, and for example, is data related to a specific concept of an academic classification method of behavior modification methods. Different behavior characteristics are set for each first agent 51 according to different behavior characteristic data 40. Further, the agent generation unit 111 generates a second agent 61 having a predetermined attribute based on the attribute data 41.

[0050] Then, the first agent processing unit 112 acts the same behavior modification content 52 on each of the plurality of first agents 51. The first agent processing unit 112 acquires reaction information 53 as a plurality of reactions obtained from each first agent 51 as a result of the action.

[0051] Next, the second agent processing unit 113 causes the second agent 61 to evaluate the behavior modification content 52 based on the reaction information 53. The output unit 115 outputs an evaluation result 70 of the behavior modification content 52.

[0052] In this way, based on the results of a plurality of reactions to the behavior modification content 52 obtained from each of the plurality of first agents 51 having different behavior characteristics, the second agent 61 evaluates the behavior modification content 52. Thereby, the evaluation of the behavior modification content 52 can be efficiently performed without using a group of real-world humans.

[0053] <Example> Next, more specific examples will be described. Hereinafter, a case where behavior characteristics are defined based on factors defined by the integrated behavior model (hereinafter referred to as "IBM factors") will be exemplified.

[0054] FIG. 6 is a diagram showing an example of a nudge message according to this embodiment. As shown in FIG. 6, a nudge message is associated with each of a plurality of IBM factors. The plurality of IBM factors may include, for example, at least two of experiential attitude, instrumental attitude, injunctive norm, descriptive norm, perceived control, self-efficacy, knowledge, skills, salience of the behavior, environmental constraints, and habit.

[0055] Note that each IBM factor shown in FIG. 6 is merely an example and is not limited to what is shown. At least two IBM factors shown in FIG. 6 may be included and defined as a higher-level IBM factor. For example, the above experiential attitude and instrumental attitude may be included in "Attitude". Alternatively, one IBM factor shown in FIG. 6 may be divided and a plurality of lower-level IBM factors may be defined. The nudge message may be associated with IBM factors at any level.

[0056] As shown in FIG. 6, each of the plurality of IBM factors may be associated with an identifier of the IBM factor (hereinafter referred to as "IBM ID") and a nudge message. In FIG. 6, for example, a nudge message assuming "residents of ○○ City" as the target person and "recommendation to undergo a specific health examination (hereinafter referred to as "specific health check")" as the behavior change is shown. The nudge message associated with each IBM factor is generated to encourage the behavior change of the target person in consideration of each IBM factor. In FIG. 6, one nudge message is associated with each IBM factor, but a plurality of nudge messages may be associated with each IBM factor. Also, the IBM ID for identifying each IBM factor is not limited to what is shown in the figure. The generated nudge message may be stored in the storage unit 100.

[0057] <Agent composition> In the present embodiment, the agent generation unit 111 defines the behavior characteristic data 40 based on the IBM factors and generates a plurality of first agents 51. The agent generation unit 111 may assign a first agent ID to each generated first agent 51.

[0058] Also, the agent generation unit 111 generates at least one second agent 61 having different roles based on the predetermined attribute data 41.

[0059] FIG. 7 is a diagram showing an example of the behavior characteristic data 40 according to the present embodiment. For example, in FIG. 7, for each first agent 51, the behavior characteristic data 40 in which the first agent ID and the parameters of each IBM factor are associated is illustrated. The parameters assigned to each IBM factor are set in the range of 1 to 10, and the larger the numerical value of the parameter, the higher the contribution rate of the IBM factor in the behavior change, and the more likely the IBM factor is to be reflected in the reaction result to the nudge message. Note that the behavior characteristic data 40 is not limited to that shown in FIG. 7.

[0060] FIG. 8 is a diagram showing an example of the attribute data 41 according to the present embodiment. For example, in FIG. 8, for each second agent 61, the attribute data 41 in which the second agent ID, the persona as a typical evaluator image, and the parameters of each IBM factor are associated with each other is illustrated. Note that the attribute data 41 is not limited to that shown in FIG. 8. For example, information regarding the attributes of the second agent 61, such as the age and gender of the second agent 61, may be added.

[0061] <Crowd AI> The first agent processing unit 112 acquires each reaction of the crowd AI 50 as reaction information 53. Here, the reaction information 53 is, for example, information indicating a reaction from the crowd AI 50 to which a nudge message is attached, and may be an answer to the nudge message, an answer regarding the situation of behavior modification (for example, the reservation status of a specific health check), an answer regarding the result of behavior modification (for example, the examination result of a specific health check), or the like.

[0062] When the first agent processing unit 112 determines that the reaction of a certain first agent 51 has ended, the first agent processing unit 112 may acquire, including in the reaction information 53, the total time from when the first agent 51 starts the reaction to when it ends, the total number of questions consumed, and the total number of information amounts.

[0063] Note that the fact that the reaction of the first agent 51 has ended may be determined, for example, when a final question is prepared in the behavior modification content 52 and the first agent processing unit 112 determines that a certain first agent 51 has reacted to the final question. Further, the first agent processing unit 112 may include a character string "This is the last question" in the behavior modification content 52 and determine the end based on the character string.

[0064] <Summary AI> The agent generation unit 111 generates the second agent 61 to be used by the second agent processing unit 113 in the generation AI system. The number of second agents 61 is arbitrary.

[0065] Based on the reaction information 53, the second agent processing unit 113 causes the second agent 61 to evaluate the action modification content 52 that has acted on the crowd AI 50 along, for example, the following evaluation axes. · The number of first agents in the crowd AI 50 that have completed the target action · The total amount of time spent by the crowd AI 50 · The amount of time or consumption information required by the latest first agent 51 among the crowd AI 50 · The difference in the amount of time or consumption information required by the fastest first agent 51 and the latest first agent 51 among the crowd AI 50

[0066] Further, when there are a plurality of second agents 61, the second agent processing unit 113 may cause each second agent 61 to evaluate along different evaluation axes. For example, the second agent 61A of a certain second agent 61 may be caused to evaluate the number of agents that have completed the target action. Also, the second agent 61B of a different second agent 61 may be caused to evaluate the difference in the amount of time or consumption information required by the fastest and the latest first agents 51 among the group of first agents 51. Further, the second agent 61C of a different second agent 61 may be caused to comprehensively evaluate the respective merits and demerits along the evaluation axis.

[0067] Further, the second agent processing unit 113 may analyze, for example, the presence or absence of a reaction to the nudge message, and compare the passages with a high reaction rate and those with a low reaction rate to evaluate the validity of the passages used in the nudge message. <Flow of processing> Next, an example of the processing executed by the information processing apparatus 10 will be described with reference to the flowchart shown in FIG. 9.

[0068] In step S101, the agent generation unit 111 generates each first agent 51 having different action characteristics, that is, the crowd AI 50.

[0069] In step S102, the first agent processing unit 112 causes the same action modification content 52 to act on each of the first agents 51 that make up the crowd AI 50, and as a result of the action, acquires reaction information 53 obtained from each of them.

[0070] In step S103, the agent generation unit 111 generates a second agent 61 having a predetermined attribute, that is, a summary AI 60.

[0071] In step S104, the second agent processing unit 113 causes the summary AI 60 to evaluate the action modification content 52 based on the reaction information 53 obtained from the crowd AI 50.

[0072] In step S105, the output unit 115 outputs the evaluation result 70 of the action modification content 52.

[0073] As described above, in the present embodiment, the second agent 61 evaluates the reaction information 53 obtained from the plurality of first agents 51 having different action characteristics, thereby enabling the evaluation of the action modification content 52. As a result, it is possible to reduce a predetermined cost such as time and labor without using a group of people in the real world.

[0074] For example, on the information processing system 1, as the action modification content 52, a questionnaire regarding the sense of aversion to vaccination is adopted, and the crowd AI 50 is made to react to the action modification content 52. Then, the summary AI 60 can be utilized to improve the architecture of the action modification content 52 by evaluating the time budget and the consciousness load budget.

[0075] Also, on the information processing system 1, a simulation space imitating human society is generated, the crowd AI 50 is arranged, and the action modification content 52 is made to react based on the reflected action modification. Then, by evaluating the reaction information 53 obtained by the summary AI 60, the action modification content 52 can be evaluated more efficiently.

[0076] Note that each of the plurality of first agents 51 belonging to the crowd AI 50 may be an agent that reflects, without overlapping with each other, a plurality of independent concepts included in a predetermined classification method related to the action variation method, or may be an agent that comprehensively reflects them.

[0077] From the above, each of the plurality of first agents 51 can generate the crowd AI 50 that comprehensively mimics specific concepts of the academic classification method of the action variation method without overlapping. As a result, by reacting the action variation content 52 to the crowd AI 50, reaction information 53 can be obtained comprehensively from each first agent 51 without overlapping. In addition, the summary AI 60 can perform an evaluation considering even a small number of reactions.

[0078] The information processing apparatus 10 includes a reception unit 114 that receives a predetermined designation content from a user, and the second agent 61 may be an agent that reflects the predetermined designation content from the user received by the reception unit 114.

[0079] For example, the reception unit 114 receives an instruction from the user to request a more detailed analysis result for the evaluation result 70 of the summary AI 60. The second agent processing unit 113 acquires the instruction, and the summary AI 60 may perform re-evaluation in accordance with the user's instruction. Also, the evaluation method and output method of the summary AI 60 may be changed according to an instruction from the user. For example, when the user instructs the summary AI 60 to calculate the total amount of time spent by a specific first agent 51, the second agent processing unit 113 may calculate the content.

[0080] In addition, the reception unit 114 receives an instruction from the user regarding the attribute of the second agent 61, and according to the instruction, the agent generation unit 111 may reflect the received attribute on the second agent 61. Note that the instruction received by the reception unit 114 may be performed, for example, by inputting an instruction related to the configuration as a prompt.

[0081] As described above, according to such a configuration, by receiving an instruction from a user and reflecting the instruction content in the crowd AI 50 and the summary AI 60, it is possible to implement the content required by the user and enhance the sense of satisfaction.

[0082] Note that this embodiment is for facilitating the understanding of the present invention and is not for limiting and interpreting the present invention. The present invention can be changed / improved without departing from its gist, and equivalents thereof are also included in the present invention. interpret. The present invention can be changed / improved without departing from its gist, and equivalents thereof are also included in the present invention. obtained, and equivalents thereof are also included in the present invention.

[0083] In the present invention, "part" does not simply mean a physical means, but also includes the case where the function of the " part" is realized by software. Also, even if the function of one "part" or device is realized by two or more physical means, devices, or software, even if the functions of two or more "parts" or devices are realized by one physical means, device, or software, it may be realized.

[0084] <Modification Example> The above embodiment or each example is for facilitating the understanding of the present invention and is not for limiting and interpreting the present invention. The present invention can be changed / improved without departing from its gist, and equivalents thereof are also included in the present invention. Also, the present invention can form various disclosures by appropriately combining a plurality of components disclosed in the above embodiment or each example. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, components may be appropriately combined from different embodiments. For example, in this example, on the generation AI system, agents are generated, etc., but the method of generating agents is not limited to the generation AI system. Various proposals have been made for AI that can generate text, etc., based on predetermined information, and the agent generation unit 111 in this embodiment is not limited to known AI that can generate text.

[0085] Note that the AI can be either a machine learning type or a non-machine learning type (rule-based). Also, the AI that generates the agent may be possessed by the information processing device 10, or as in this embodiment, a generation AI system provided by an external system may be used. As the generation AI system provided by the external system, for example, generation AI systems such as ChatGPT, Google Bard (registered trademark), Claude (registered trademark), etc. may be used.

[0086] <Behavior characteristic data and other cases> In the modification example, as a specific concept of the academic classification method of the behavior modification method, the Health Belief Model (HBM) will be described as an example. The Health Belief Model is one of the health behavior theories, and the main factors that increase the likelihood of a person performing health-promoting behaviors are the recognition of threats and the balance between benefits and drawbacks.

[0087] The recognition of threat means feeling a sense of crisis that "this is bad" as it is. To feel such a sense of crisis, it is necessary to recognize both the following "likelihood" and "severity". The recognition of likelihood means feeling that there is a high possibility of oneself getting sick or having a complication as it is. The recognition of severity means feeling that if one gets sick or has a complication, the result is significant in terms of health, economy, society, etc. The balance between benefits and drawbacks means feeling that the benefits of performing a health-promoting behavior are greater for oneself when considering the benefits and drawbacks of performing that behavior.

[0088] The storage unit 100 may store behavior characteristic data 40 to be reflected in each agent, which is dataized for different behavior characteristics for the factors defined by the Health Belief Model (HBM).

[0089] For example, the nudge message is a message that assumes "residents of ○○ City" as the target audience and "recommendation to undergo a specific health examination (hereinafter referred to as 'Specific Health Check')" as the behavior change. The nudge message associated with the factors defined in each Health Belief Model (HBM) is generated to encourage the behavior change of the target audience in consideration of the factors defined in each Health Belief Model (HBM). The storage unit 100 may store the nudge message generated in advance.

[0090] For example, it is assumed that the target audience is "residents of ○○ City" and the behavior change is "undergoing a specific health check (e.g., a medical examination)", but it is not limited to this. The target audience is not limited to users of administrative services and may be users of various services such as English conversation and qualification exams. Also, the behavior change is not limited to the use of administrative services and may be continuous learning, use of various services, etc. <Behavior characteristic data BCTTv1>

[0091] The behavior change content 52, the behavior characteristic data 40 and / or the attribute data 41 may be generated, for example, based on a behavior change method (BCT: behavior change technique). For example, according to BCTTv1 (Michie S, Richardson M, Johnston M, et al.: The behavior change technique taxonomy (v1) of 93 hierarchically clustered techniques: building an international consensus for the reporting of behavior change interventions. Ann Behav Med 2013; 46: 81~95.), 93 BCTs in 16 groups are defined. Note that the definition of BCT is not limited to BCTTv1 and may be defined in any way as long as it comprehensively covers the methods of behavior change.

[0092] BCTTv1 defines 16 BCT groups: "1. Goals and planning", "2. Feedback and monitoring", "3. Social support", "4. Shaping knowledge", "5. Natural consequences", "6. Comparison of behaviour", "7. Associations", "8. Repetition and substitution", "9. Comparison of outcomes", "10. Reward and threat", "11. Regulation", "12. Antecedents", "13. Identity", "14. Scheduled consequences", "15. Self-belief", and "16. Covert learning".

[0093] Figure 10 shows an example of the classification of BCTs. In BCTTv1, each of the 16 BCT groups shown in Figure 10 has one or more BCTs belonging to it. For example, the BCT group "5. Natural consequences" includes BCTs such as "5.5. Anticipated regret". Also, the BCT group "10. Reward and threat" includes BCTs such as "10.11. Future punishment". Although not shown, one or more BCTs also belong to other groups.

[0094] Moreover, each BCT has components, and the degree of each BCT included in a certain content may be shown as a component value. Also, the sum of the component values of each BCT belonging to the same BCT group may be shown as the component value of the BCT group.

[0095] The behavioral changes prompted by the behavioral change content 52, the behavioral characteristic data 40, and / or the attribute data 41 are assumed to include, for example, language learning, dieting, purchasing financial products offered by financial institutions, undergoing regular medical check-ups, and using public services, but are not limited thereto. Also, the behavioral change content 52 is not limited to being provided to the user by electromagnetic means such as applications installed on the terminal, web pages, e-mails, short messages, etc., and may be provided to the user using other methods (e.g., customer service, mail, etc.). Further, the behavioral change content 52 is, for example, text information regarding a specific service (e.g., chat logs, etc.), moving images, still images, voices of conversations, data regarding applications, etc., but is not limited thereto, and may be any information regarding the content of the behavioral change.

[0096] In generating and evaluating the behavioral change content 52, the behavioral characteristic data 40, and / or the attribute data 41, a coordinate system with multiple behavioral change factors as axes is generated, and a target coordinate or a target area (hereinafter, these are collectively referred to as the target area) is set in this coordinate system. Then, whether the user has made a behavioral change may be evaluated in the coordinate system by whether the user's coordinates have approached the target area. Hereinafter, the distance calculated in this coordinate system is referred to as the "Behavioral Scientific Distance (BSD)".

[0097] In such a coordinate system, for each user, by setting the current coordinates and the target area that the intervention should aim for, the behavioral scientific distance from the current coordinates to the target area can be expressed as a multi-dimensional vector. Also, it becomes possible to objectively and clearly visualize the BCT to be adopted to fill this distance. That is, higher-quality behavioral change content 52 is, for example, behavioral change content 52 that reduces the behavioral scientific distance from the current coordinates to the target area.

[0098] Also, the behavior modification factors may be factors for causing a user to perform a target behavior, for example. Based on surveys such as academic surveys or awareness surveys, theories related to behavioral science, user persona settings, or behavioral process maps, etc., a plurality of behavior modification factors in the target field can be identified. Examples of behavior modification factors include "Capacity", "Opportunity", and "Motivation" in the COM-B model commonly used in behavioral science.

[0099] Note that each of the above IBM factors is merely an example, and higher-level factors including the above IBM factors may be specified as behavior modification factors. For example, the above experiential attitude and instrumental attitude may be included in "Attitude". The above injunctive norm and descriptive norm may be included in "Perceived norm". The above sense of behavioral control and self-efficacy may be included in "Personal Agency". The above knowledge and technology may be included in "Knowledge". The importance of behavior may be included in "Importance", and environmental constraints may be included in "Friction". Alternatively, lower-level factors obtained by dividing the above IBM factors may be specified as behavior modification factors. In the present embodiment, any factors related to human behavior will, such as factors defined by behavioral models other than the COM-B model and IBM, may be adopted as behavior modification factors.

[0100] The ease of response to behavior modification factors can vary from user to user. That is, in one user, behavior modification is likely to occur due to behavior modification content 52 that strongly acts on the first behavior modification factor, while in another user, behavior modification may be likely to occur due to behavior modification content 52 that strongly acts on the second behavior modification factor. Also, even for the same user, the ease of response to behavior modification factors can vary depending on the passage of time, the situation such as the time when the behavior modification content 52 is given and the environment at that time, and the target behavior and the content of the behavior modification. The ease of response to behavior modification factors may be expressed as behavior characteristics of the user. Thus, the behavioral science distance at which the behavior modification content 52 can act can vary depending on the behavior characteristics of the user.

Explanation of Signs

[0101] 1… Information processing system, 10… Information processing device, 11… Processor, 12… Storage device, 13… Communication IF, 14… Input device, 15… Output device, 20… Information processing terminal, 40… Behavior characteristic data, 41… Attribute data, 50… Crowd AI, 51… First agent, 52… Behavior modification content, 53… Response information, 60… Summarizing AI, 61… Second agent, 70… Evaluation result, 100… Storage unit, 110… Control unit, 111… Agent generation unit, 112… First agent processing unit, 113… Second agent processing unit, 114… Reception unit, 115… Output unit, 200… Storage unit, 201… Communication unit, 202… UI unit, 203… Control unit

Claims

1. a first agent processing unit that acquires a plurality of responses obtained from each of a plurality of first agents having different behavioral characteristics as a result of applying the same behavior modification content to each of the plurality of first agents; a second agent processing unit that causes a second agent having a predetermined attribute to evaluate the behavior change content based on a plurality of responses obtained from each of the plurality of first agents; an output unit that outputs an evaluation result of the behavior change content; An information processing device comprising:

2. Each of the plurality of first agents is an agent that reflects a plurality of independent concepts included in a predetermined taxonomy related to behavior modification techniques without overlapping with each other. The information processing device according to claim 1 .

3. Each of the plurality of first agents is an agent that comprehensively reflects a plurality of independent concepts included in a predetermined taxonomy regarding behavior modification techniques. The information processing device according to claim 1 .

4. The device further includes a reception unit for receiving predetermined designation content from a user, the second agent is an agent that reflects the predetermined designation content received from the user; The information processing device according to claim 1 .

5. an acquisition step of acquiring a plurality of responses obtained from each of a plurality of first agents having a plurality of different behavioral characteristics as a result of applying the same behavior modification content to each of the plurality of first agents; an evaluation step of having a second agent having a predetermined attribute evaluate the behavior change content based on a plurality of responses obtained from each of the plurality of first agents; and outputting the evaluation result of the behavior change content. An information processing method executed by an information processing device.

6. an acquisition step of acquiring a plurality of responses obtained from each of a plurality of first agents having a plurality of different behavioral characteristics as a result of applying the same behavior modification content to each of the plurality of first agents; an evaluation step of having a second agent having a predetermined attribute evaluate the behavior change content based on a plurality of responses obtained from each of the plurality of first agents; an output step of outputting a result of the evaluation of the behavior change content; A program for causing a computer to execute the following.

Images