Behavior modification inducing system and behavior modification inducing method

Abstract

To design an intervention measure for guiding an action of an opponent to be changed into an action (target action) preferable for the opponent.SOLUTION: A biological change inducing system (300) includes means (131A) for reproducing a thinking pattern of a target subject using a prediction model machine-learned using biological information of the target subject as teacher data, means (131B) for predicting a reaction of the target subject to a current external stimulus to the target subject on the basis of the thinking pattern of the target subject using the current external stimulus as an input, and means (131C) for designing an interventional measure for inducing an action of the target subject to a target action.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a behavioral change induction system, a behavioral change induction method, and a program for causing a computer to execute the method, which uses machine learning to learn a target subject's thought patterns and behavioral patterns from the target subject's past reactions (behavior) when exposed to external stimuli, predicts the reaction (behavior) that the target subject will have when exposed to current external stimuli or external stimuli expected in the near future, and designs intervention measures to induce the target subject to change their behavior to a target behavior (desirable behavior) if the reaction is not desirable. [Background technology]

[0002] The United States Department of Defense (Pentagon) published a document entitled "Psychological Operations" ("Psychological Warfare") (dated January 7, 2010) (Non-Patent Document 1). Within the concept of "Psychological Operations" (hereafter referred to as "PsyOp") as defined in this document, the spread of disinformation by "State Players" sponsored by the military or government has been recognized as a core PsyOp or as one of the new tactical weapons. By spreading false information, it is possible to mislead the decisions and actions of the opposing government and military, thereby supporting one's own country's own physical attacks, such as missile attacks. In this way, the spread of false information is being positioned as a new tactical weapon to replace physical attacks.

[0003] In today's world, where we have moved from the age of PsyOps to the age of cognitive warfare, information (including disinformation) is a strategic weapon aimed at changing the perceptions of each individual citizen, including the enemy's top brass, to suit the goals of one's own country, and is being used to attack the perceptions of each individual, with the aim of shaping public opinion around the world and targeting the civilian population of enemy countries. PsyOps is primarily concerned with the spread of disinformation as a tactical weapon in existing warfare domains such as land, sea, and air, but in modern cognitive warfare, the cognitive domain itself has become a warfare domain, and in addition to tactical aspects such as disrupting and destroying the enemy's leadership, it is also necessary to understand its role as a strategic weapon targeting the cognitive transformation of individual civilians. This is similar to how cyber technology, which was once an offensive technology supporting physical tactical weapons in traditional warfare domains, has now become a warfare domain itself.

[0004] In modern cognitive warfare, the information generated and utilized goes beyond simply disrupting the enemy country's chain of command; it is produced in an integrated manner based on some medium- to long-term benefit of one's own country, and in many cases, large amounts of information generated by AI, such as generative AI, are highly combined and disseminated. In modern times, not only language and other symbolic information is used, but also quasi-symbolic information such as subliminal messages. Furthermore, non-symbolic brain information processing such as biofeedback is becoming possible to directly intervene in the thoughts of the target subject and rewrite the information. To utilize these technologies or prevent them, it is necessary to abstract all patterns from large amounts of information at different levels of abstraction, such as the five senses and language, recognize the enemy's intentions, and utilize the information. This requires multiple AIs to operate in a cooperative and distributed manner and perform inference and analysis at all levels of abstraction. [Prior art documents] [Non-patent literature]

[0005] [Non-Patent Document 1] https: / / irp.fas.org / doddir / dod / jp3-13-2.pdf Summary of the Invention [Problem to be solved by the invention]

[0006] As mentioned above, in modern times, information, whether true or false, has become a strategic weapon aimed at changing the perceptions of each and every citizen, including the enemy country's top brass, to suit the goals of one's own country. Furthermore, handling information in a way that suits one's own goals is an extremely important strategy not only in the field of strategic weapons but also in the business field. The same can be said in the everyday lives of individuals. However, in the fields of strategy and business, the use of information in a way that is consistent with such strategies has not yet been fully developed. The present invention has been made in view of the above circumstances, and aims to provide a system and method for utilizing information in a manner that is consistent with the above strategies. As there is no prior art similar to the present invention, the above-mentioned document on PsyOp published by the Pentagon is cited as a prior art document. [Means for solving the problem]

[0007] To achieve this object, the present invention provides, in a first aspect, a behavior change induction system (300) for designing an intervention plan for inducing behavior change in a target subject (172), comprising: a storage means (132) for storing biometric information including the target subject's (172) reaction to past external stimuli of the target subject (172); a thought pattern reproduction means (131A) for reproducing the thought pattern or behavior pattern of the target subject (172) using a prediction model (210) machine-learned using the biometric information as training data (200); and a thought pattern reproduction means (131B) for reproducing the thought pattern or behavior pattern of the target subject (172) in response to current external stimuli or behavior of the target subject (172). and a reaction estimation means (131B) that receives as input current or predicted future external stimuli and outputs a predicted reaction of the target subject (172) to the external stimuli based on the thought pattern or behavior pattern of the target subject (172) reproduced by the thought pattern reproduction means (131A), and an intervention design means (131C) that designs an intervention to induce the behavior of the target subject (172) to a target behavior in accordance with the predicted reaction of the target subject (172) output by the reaction estimation means (131B).

[0008] Preferably, the biometric information includes brain pattern information of the target subject (172). For example, it is preferable that the intervention plan design means (131C) designs at least one of an intervention plan for inducing the target subject (172) to weaken the emotion caused by the external stimulus and an intervention plan for inducing the target subject (172) to replace the emotion caused by the external stimulus with another emotion. In a second aspect, the present invention provides a portable wireless communication device (100) incorporating the above-described behavioral change induction system (300).

[0009] In a third aspect, the present invention provides a behavior change promotion system for designing and implementing an intervention to induce behavior change in a target subject (172), the behavior change promotion system comprising a behavior change promotion system (300) and a portable wireless communication device (173) possessed by the target subject (172), the behavior change promotion system (300) comprising: a storage means (132) for storing biometric information including the target subject's (172) reaction to past external stimuli to the target subject (172); a thought pattern reproduction means (131A) for reproducing the thought pattern or behavior pattern of the target subject (172) using a prediction model (210) machine-learned using the biometric information as training data (200); and a thought pattern reproduction means (131A) for reproducing the thought pattern or behavior pattern of the target subject (172) reproduced by the thought pattern reproduction means (131A) using a current external stimuli or a predicted future external stimuli to the target subject (172) as an input. and an intervention plan design means (131C) for designing an intervention plan for guiding the behavior of the target subject (172) to a target behavior in accordance with the predicted reaction of the target subject (172) output by the reaction estimation means (131B), wherein the portable wireless communication device (173) has a built-in application (174) for designing the intervention plan based on the intervention plan, the behavior change induction system (100) transmits the intervention plan to the portable wireless communication device (173) of the target subject (172), and the application (174) of the portable wireless communication device (173) that receives the intervention plan formulates an intervention action customized for each target subject (172) based on the intervention plan.

[0010] In a fourth aspect, the present invention provides a method for inducing behavioral change in a target subject (172) by designing an intervention plan for inducing behavioral change, the method comprising: a first step (S110) of storing biometric information including the target subject's (172) reaction to past external stimuli; a second step (S130) of reproducing the target subject's (172) thought patterns or behavioral patterns using a prediction model (210) machine-learned using the biometric information as training data (200); and a second step (S130) of reproducing the target subject's (172) thought patterns or behavioral patterns using a prediction model (210) machine-learned using the biometric information as training data (200). The present invention provides a method for inducing behavioral change, which comprises a third process (S140) that takes as input future external stimuli to be predicted and outputs an expected reaction of the target subject (172) to those external stimuli based on the thought patterns or behavior patterns of the target subject (172) reproduced in the second process (S130), and a fourth process (S160) that designs an intervention plan for inducing the behavior of the target subject (172) to a target behavior in accordance with the expected reaction of the target subject (172) output in the third process (S140).

[0011] The fourth step (S160) preferably involves designing at least one of an intervention plan for inducing the target subject (172) to weaken the emotion caused by the external stimulus and an intervention plan for inducing the target subject (172) to replace the emotion caused by the external stimulus with another emotion. In a fifth aspect, the present invention provides a behavior change induction method for designing an intervention plan for inducing behavior change in a target subject (172), the method comprising: a first step (S110) of storing biometric information including the target subject's (172) reaction to past external stimuli; a second step (S130) of reproducing the target subject's (172) thought patterns or behavior patterns using a prediction model (210) machine-learned using the biometric information as training data (200); and a second step (S130) of inputting current external stimuli or predicted future external stimuli to the target subject (172) and reproducing the target subject's (172) thought patterns or behavior patterns reproduced in the second step (S130). The method for inducing behavioral change includes a third process (S140) for outputting an expected reaction of the target subject (172) to those external stimuli based on the thought pattern or behavior pattern; a fourth process (S160) for designing an intervention policy of an intervention measure for inducing the behavior of the target subject (172) to a target behavior in accordance with the expected reaction of the target subject (172) output in the third process (S140); and a fifth process (S170) for an application built in a portable wireless communication device held by the target subject (172) to formulate an intervention action customized for each target subject (172) based on the intervention policy.

[0012] The fourth step (S160) preferably involves designing at least one of an intervention plan for inducing the target subject (172) to weaken the emotion caused by the external stimulus and an intervention plan for inducing the target subject (172) to replace the emotion caused by the external stimulus with another emotion. In a sixth aspect, the present invention provides a program (131A, 131B, 131C) for causing a computer to execute the above-described behavioral modification induction method. The reference numerals in parentheses are used to indicate correspondence with the embodiments described later, and are not intended to limit the scope of the claims.

[0013] As described above, the present invention aims to realize behavioral changes in target subjects. For example, when this invention is applied to the military field, one example would be to make the leader of an enemy country feel that "our country is in an advantageous position" and make him decide to avoid starting a war against our country, or to make the leader of an enemy country unconsciously choose an action that will give our country an advantage, causing him to feel a sense of failure that "the strategy backfired." The application field of the present invention is not limited to the military field, but can also be applied to the business field. For example, by implementing the present invention in the marketing field, it is possible to encourage consumers to purchase one's own products. Existing behavior change approaches (methods for changing the behavior of others so that they behave in a way that one desires), particularly behaviorist approaches, have mainly relied on the "event-driven model," which predicts future events based on past events. This event-driven model requires time-series information on events to go through a process of inferring output events that are correlated with specific input events, and has the following fundamental limitations:

[0014] Event-driven models only react based on input events. That is, they are reactive in nature. This means they can only respond to known patterns and predictable events, not proactively. For example, in military operations, it is inevitable that a country will always be at a disadvantage because it will be reactive to the enemy. Event-driven models rely solely on past data, making them inflexible and unable to respond appropriately to unknown or unexpected events. For example, when it comes to unprecedented events such as a new virus pandemic or the emergence of unknown technologies, it is impossible to formulate countermeasures using only past data. This problem is particularly pronounced in the fields of military and security, where the ability to respond immediately to new and unpredictable threats is required. Furthermore, in previous behavioral approaches, people's behavioral tendencies were abstracted to a certain extent and treated as a whole, making it impossible to address the behavioral tendencies of each individual.

[0015] Inference systems such as general large-scale language models (LLMs) use a wide range of text, audio, image, and video information as input data, without targeting a specific person. This data is often fictional or emotional descriptions that do not necessarily accurately reflect actual human thought patterns. Therefore, inference systems built on this general data do not necessarily accurately represent the thought or behavior patterns of a specific individual. To solve these problems, the present invention employs an approach that does not rely on time-series event data but is based on direct analysis of target subjects' biological information (e.g., brain patterns obtained by brain imaging devices).

[0016] As will be described later, this approach uses three types of inference systems with different roles (thought reproduction AI, behavior prediction AI, and behavior change AI), each of which analyzes the individual biometric information (brain patterns, etc.) of the target subject, predicts the target subject's future behavior, and designs specific intervention measures for those behaviors. Furthermore, in the present invention, as will be described later, multiple inference systems of the same type are operated in parallel in order to reproduce the target subject's cognitive model in more detail. This parallelization makes it possible to reproduce the multiple thought patterns inherent in the target subject and make optimal inferences tailored to the situation. It also makes it possible to predict the possibility of behavioral change in the target subject in real time and quickly design effective intervention measures. Furthermore, by using a simulation-based behavior prediction system to simulate in advance the intervention measures for behavior change proposed by the above-mentioned inference system and repeating the verification and evaluation, it is possible to create optimal intervention measures.

[0017] This system is particularly valuable in military situations where rapid decision-making is required and the battle situation can change in an instant, making it possible to predict enemy attack intentions in advance and plan and execute rapid defensive actions. As described above, the present invention is based on an approach that directly analyzes the biometric information of target subjects. Specifically, the present invention breaks away from the conventional constraints of relying on time-series event data and instead adopts a cognitive model architecture based on cognitive science to achieve the goal of "whom we want to encourage, when, and what kind of behavior we want to encourage." [Effects of the Invention]

[0018] According to the present invention, it is possible to analyze the brain patterns and other biological information of a target subject in real time, predict the target subject's behavior before an event occurs, and guide the target subject to change their behavior, thereby enabling strategic countermeasures to be taken before the target subject is affected, overcoming the conventional reactive, after-the-fact approach. In the present invention, behavior prediction is performed directly from current biological information without relying on past event patterns, so it is possible to quickly adapt to new events and changes and provide countermeasures. In the present invention, detailed biometric information of each target subject is directly analyzed, making it possible to predict the individual behavior of each target subject, thereby more accurately understanding each target subject's actual cognitive patterns and behavioral tendencies and creating more precisely personalized interventions for each target subject.

[0019] According to the present invention, an intervention action customized for each target person can be generated, thereby improving the effectiveness of the intervention for each target person. Furthermore, in the process of intervention, it is possible to divide the roles between cloud computers and edge computers, reducing the use of computing resources, which allows large-scale campaigns and information dissemination activities to be carried out efficiently with limited resources. [Brief explanation of the drawings]

[0020] [Figure 1] 1 is a block diagram showing an example of the structure of a behavior modification induction system according to a first embodiment of the present invention. [Figure 2] FIG. 2 is a conceptual diagram showing a configuration of a first program. [Figure 3] FIG. 2 is a conceptual diagram showing functions of first to third programs. [Figure 4] 2 is a flowchart showing the operation of the behavior modification induction system shown in FIG. 1. [Figure 5] FIG. 10 is a conceptual diagram showing the cooperative relationship between the first to third programs. [Figure 6] This is a conceptual diagram of how the first program (thought reproduction AI) builds and maintains a cognitive model. [Figure 7] FIG. 10 is a diagram illustrating an example of biometric information of a target subject. [Figure 8] FIG. 10 is a diagram illustrating an example of behavioral information of a target subject. [Figure 9] FIG. 10 is a diagram showing an example of stimulus information that a target subject receives from the external environment. [Figure 10] FIG. 10 is a conceptual diagram of a behavior prediction system configured by a second program.

[0021] [Figure 11] This is a conceptual diagram of a behavior change system composed of the third program. [Figure 12] FIG. 1 is a schematic diagram of the intervention delivery system. [Figure 13] FIG. 1 is a diagram showing an example of inputs and outputs (intervention policy signals and intervention actions) in the security domain. [Figure 14]FIG. 1 is a diagram showing an example of inputs and outputs (intervention policy signals and intervention actions) in the marketing domain. [Figure 15] FIG. 1 is a diagram showing an example of inputs and outputs (intervention policy signals and intervention actions) in the entertainment domain. DETAILED DESCRIPTION OF THE INVENTION

[0022] FIG. 1 is a block diagram showing an example of the structure of a portable wireless communication device 100 equipped with a behavioral change induction system according to a first embodiment of the present invention. The portable wireless communication device 100 includes a communication unit 110, a control unit 120, an external memory (hard disk) 130, an input / output unit 140, and an antenna 150. Of these, the behavioral modification induction system 300 is composed of a control unit 120 and an external memory (hard disk) 130. The communication unit 110 is connected to an antenna 150, and transmits and receives data via the antenna 150 to and from other wireless communication devices via wireless communication. The communication unit 110 includes a wireless receiving unit 111 , a wireless transmitting unit 112 , and a changeover switch 113 .

[0023] The wireless receiving unit 111 demodulates data received from other wireless communication devices and sends the data to the control unit 120. The wireless transmitting unit 112 modulates the data output from the control unit 120 and transmits the data to other mobile phone devices or other wireless communication devices via the antenna 150. The changeover switch 113 receives a signal from the control unit 120 and switches between transmission (modulation) and reception (demodulation) in accordance with the signal. The control unit 120 is composed of a central processing unit (CPU) 121, a first memory 122 consisting of ROM, a second memory 123 consisting of RAM, an input interface 124 for transferring various commands and data input to the control unit 120 to the central processing unit 121, an output interface 125 for outputting the processing results executed by the central processing unit 121 to the outside, and a bus 126 connecting the central processing unit 121 to each of the first memory 122, the second memory 123, the input interface 124 and the output interface 125.

[0024] The first memory 122 stores various control programs executed by the central processing unit 121 and other non-rewritable data. The second memory 123 stores various data and parameters and provides a working area for the central processing unit 121. In other words, data or programs temporarily required for the central processing unit 121 to execute various control programs are read from the external memory 130 and temporarily stored in the second memory 123. The central processing unit 121, together with an OS (Operating System), controls the overall operation of the portable wireless communication device 100. Specifically, it reads first to third programs 131A, 131B, and 131C (described later) from the external memory 130 and executes these programs. That is, the central processing unit 121 operates in accordance with the first to third programs 131A, 131B, and 131C stored in the external memory 130. As will be described later, the central processing unit 121 outputs a predetermined result in response to a predetermined input via a trained evaluation model.

[0025] The input / output unit 140 is made up of an operation unit 141, a display 142, a speaker 143, and a microphone 144. The operation unit 141 is made up of, for example, a numeric keypad, and various data are input to the control unit 120 via the operation unit 141 . The display 142 is, for example, a liquid crystal display, and displays the results of calculations performed by the control unit 120 and other data on the screen. Audio data received from the outside is output from speaker 143 via wireless receiving unit 111 of communication unit 110. The voice of the user of portable wireless communication device 100 is collected via microphone 144 and output to the outside via wireless transmitting unit 112 of communication unit 110. The external memory (hard disk) 130 is composed of an application section 131 and a data storage section 132 .

[0026] The data storage unit 132 is composed of a biometric information storage unit 132A that stores the brain patterns and other biometric information of the target subject, which will be described later, and a sub-data storage unit 132B that stores various data related to the target subject other than the biometric information. The application unit 131 stores an OS 131S that controls the overall operation of the portable wireless communication device 100, a first program 131A, a second program 131B, and a third program 131C. The first program 131A, the second program 131B, and the third program 131C are programs that operate as a thought reproduction system, a behavior prediction system, and a behavior modification system, respectively, which will be described later. FIG. 2 is a conceptual diagram showing the configuration of the first program 131A. As shown in FIG. 2, the first program 131A is made up of a teacher data input program 135A, an evaluation model generation program 135B, and an evaluation result output program 135C.

[0027] Like the first program 131A, the second program 131B and the third program 131C are also made up of these three programs. The teacher data input program 135A is a program for inputting teacher data to the central processing unit 121 for generating an evaluation model by machine learning. The evaluation model generation program 135B generates each evaluation model by machine learning using the training data input by the training data input program 135A. As an algorithm for generating the evaluation model, for example, a neural network is used. The evaluation result output program 135C outputs the trained evaluation model generated by the evaluation model generation program 135B to the central processing unit 121.

[0028] FIG. 3 is a conceptual diagram showing the functions of the first to third programs 131A, 131B, and 131C. As shown in FIG. 3, training data 200 is sent to an evaluation model generation program 135B by a training data input program 135A, and the evaluation model generation program 135B generates an evaluation model 210 based on the training data 200. This evaluation model 210 undergoes continuous machine learning and becomes a trained evaluation model. Data 220 corresponding to the first to third programs 131A, 131B, and 131C is input to the trained evaluation model 210. For example, the first program 131A receives as input, as data 220, the biometric information of the target subject sent from the biometric information storage unit 132A and other data related to the target subject sent from the sub-data storage unit 132B. The trained evaluation model 210 generates predetermined evaluation results (such as the target subject's thought patterns, behavioral predictions of the target subject, and intervention measures, which will be described later) based on the data 220, and outputs them as output 230.

[0029] The evaluation results generated by the trained evaluation model 210 are output to the central processing unit 121 by the evaluation result output program 135C, and, if necessary, are displayed on the display 142 or printed via a printer (not shown). 4 is a flowchart showing the operation of the portable wireless communication device 100. Below, each operation performed by the portable wireless communication device 100 will be outlined with reference to FIG. First, the target subject's thought patterns, behavior patterns, and other biometric information and various information about the target subject are collected from past reactions (statements and actions) made by the target subject when exposed to external stimuli, and the biometric information is stored in the biometric information storage unit 132A of the external memory 130, and the other information is stored in the sub-data storage unit 132B (step S110). Biometric information also includes, for example, brain pattern information obtained by brain imaging devices (such as functional Magnetic Resonance Imaging (fMRI) and electroencephalogram (EEG)).

[0030] These biological information and other information are sent as training data by the training data input program 135A of the first program 131A to the evaluation model generation program 135B, which then generates the evaluation model 210 by machine learning (step S120). Furthermore, the evaluation model generation program 135B uses the machine-learned evaluation model to reproduce the thought patterns and behavior patterns of the target subject (step S130). The reproduced thought patterns and behavior patterns of the target subject are sent to the second program 131B by the evaluation result output program 135C. In the second program 131B, current external stimuli for the target subject or external stimuli expected in the near future are input to the evaluation model, and the reaction (statements and actions) that the target subject will have when exposed to these external stimuli is predicted via the evaluation model already generated by the evaluation model generation program 135B (step S140).

[0031] The reaction that the target subject is likely to take is sent to the third program 131C via the evaluation result output program 135C of the second program 131B, and the third program 131C determines whether the reaction of the target subject is favorable to the user of the portable wireless communication device 100 (step S150). If the target subject's response is favorable (YES in step S150), no intervention measures (described below) for the target subject are considered, and the operation of the portable wireless communication device 100 ends here. If the target subject's response is not favorable (NO in step S150), the evaluation model generation program 135B of the third program 131C designs an intervention to induce the target subject's behavior to change to a target behavior (behavior favorable to the user of the portable wireless communication device 100) via the evaluation model that has already been generated (step S160). Thereafter, this intervention is carried out as appropriate (step S170).

[0032] The operation of the portable wireless communication device 100 will now be described. The first program 131A stored in the application section 131 of the external memory 130 functions as a thought reproduction system that reproduces the thoughts of a target subject, and includes a thought reproduction AI (Sub-Belief System AI = SBS-AI) as a learning means. The second program 131B functions as a behavior prediction system, and includes a behavior prediction AI (Prediction-AI) as a learning means. The third program 131C functions as a behavior modification system, and includes a behavior modification AI (Behavior Modification AI) as a learning means. FIG. 5 is a conceptual diagram showing the cooperative relationship between the first to third programs 131A, 131B, and 131C.

[0033] As shown in Figure 5, in outline, the target subject's biological information (the target subject's response to external stimuli) and other information about the target subject (the target subject's behavioral information, and brain stimulation information obtained from the external environment through the five senses and language) are input to first program 131A (thought reproduction AI) as input information 220 (see Figure 3), and first program 131A (thought reproduction AI) constructs an inference system for each target subject based on this input information 220. This inference system reproduces the target subject's thought patterns or behavior patterns, and the reproduced thought patterns or behavior patterns are output 230 (see Figure 3). The second program 131B (behavior prediction AI) takes as input 220 the thought and behavior patterns of the target subject reproduced by the first program 131A, and based on these thought and behavior patterns, predicts what behavior the target subject will exhibit thereafter or after a specific time has passed, or what thought pattern (brain pattern) and behavior pattern the target subject will have, and outputs this prediction as output 230, constituting a behavior prediction system.

[0034] This behavioral prediction predicts the potential behavioral patterns of the target subject when they are in a natural state (a peacetime situation, not an emergency such as wartime or disaster). Each behavior prediction AI is constructed as a pair with each thought reproduction AI, and when the thought reproduction AIs are parallelized, the same number of behavior prediction AIs are used to construct the parallelization. The first program 131A and the second program 131B constitute a cognitive model 160 for the target subject. The third program 131C (behavior change system) receives as input 220 the behavior prediction of the target subject output by the second program 131B (behavior prediction AI), and if the future behavior of the target subject predicted by the second program 131B is not desirable, formulates an intervention (action plan) for changing the behavior of the target subject to desirable behavior (target behavior), and outputs this intervention 230. The formulated intervention is executed on the target subject in the real world, inducing a behavior change in the target subject.

[0035] To build behavioral modification AI, we use experimental data obtained from multiple subjects regarding intervention methods used in cognitive warfare. These three first to third programs 131A, 131B, and 131C function in cooperation with each other to form a cognitive-driven model that is essentially different from the conventional event-driven model. Examples of interventions developed by the third program, 131C (Behavior Change System), include a variety of methods, such as mass media, social media, verbal inducements, and liminal and subliminal effects through images and videos. In recent years, medical devices that directly affect the human body have also been developed, and methods of directly intervening in an individual's brain have been developed using advanced brain imaging devices such as functional magnetic resonance imaging (fMRI), electroencephalography (EEG), and magnetoencephalography (MEG).

[0036] As described above, in the portable wireless communication device 100, the thought reproduction AI analyzes the natural language messages of the target person and recognizes the hidden intentions of the target person. For example, the brain activation state measured in real time by fMRI (functional magnetic resonance imaging) is mutually fed back to the thought reproduction AI, and the hidden intentions can be read from the pattern and the reaction when a signal is sent to it interveningly. Behavior prediction AI can hypothesize multiple intentions, rather than just a single one. By identifying the brain activation state for each intention in advance using fMRI or other techniques, it becomes possible to predict the thoughts and actions of the target subject. Below, the first program 131A (thought reproduction AI), the second program 131B (behavior prediction AI), and the third program 131C (behavior modification system) will be described in detail.

[0037] (A) First Program 131A (Thought Reproduction System, Thought Reproduction AI) Figure 6 is a conceptual diagram of the construction and updating of cognitive models using the thought reproduction AI (SBS-AI). The first program 131A constructs a "thought reproduction AI that expresses brain patterns" as an inference system for each of the target subjects. As shown in FIG. 6, the inputs 220 (see FIG. 3) to the thought reproduction AI include the time-stamped biometric information of the target subject sent from the biometric information storage unit 132A, as well as behavioral information of the target subject sent from the sub-data storage unit 132B, and stimulus information that the target subject receives from the external environment through the five senses and language. Representative examples of input 220, detection devices, and detection data are shown in Figures 7 to 9. Figure 7 shows an example of biometric information of a target subject, Figure 8 shows an example of behavioral information of a target subject, and Figure 9 shows an example of stimulus information received by the target subject from the external environment.

[0038] The cognitive model 160 configured by the first program 131A (thought reproduction AI) has the function of numerically analyzing correlations from the input 220 (input information) and finding out the thought patterns and behavior patterns of the target subject. Numerical analysis methods include principal component analysis used in traditional statistics, support vector machines used in machine learning, deep learning (AI) statistical processing using huge amounts of data, and large-scale language models (LLM).As analytical methods are rapidly evolving, it is best to adopt the most appropriate method at any given time. In the first program 131A (thought reproduction AI), for example, a neural network is used. The thought replicating AI is updated by applying backpropagation to the thought replicating AI using continuously received input 220. This continuous feedback keeps the cognitive model 160 of the target subject up to date. There are multiple ways to obtain input 220 (input information) on an ongoing basis.

[0039] For example, data on the target subjects can be obtained through portable wireless communication devices (smartphones, wearable devices, small sensors) used by the target subjects themselves. Alternatively, data on the target audience can be obtained from IoT devices or smart home devices installed in the target audience's vicinity (home or office). Alternatively, data can be obtained by tracking the internet activity of your target audience, such as their social media usage, online search behavior, and website browsing history. Alternatively, you can collect feedback directly from your target audience through online surveys or in-app pop-up surveys.

[0040] Furthermore, by using a device equipped with biosensing technology (e.g., Apple Vision Pro (registered trademark)), it is possible to directly obtain the target subject's biological information, such as brain waves, muscle activity, and eye movement. In the real world, all information surrounding a target subject (symbolic information such as language, quasi-symbolic information such as subliminal messages, non-symbolic information such as biofeedback, etc.) causes neural feedback at various levels of abstraction, updating the target subject's cognition. This is a phenomenon called homeostatic entrainment, which arises from the interaction between internal representations and the external environment, and is innately possessed by living organisms. Therefore, the method of updating a thought reproduction AI using the backpropagation method with input 220 (input information) as result data can be said to be a logic that mimics the homeostatic entrainment phenomenon in a computer. As a result, by continuously updating the thought reproduction AI, the cognitive characteristics of the target subject himself / herself and the behavioral characteristics of the cognitive model 160 can be brought closer to each other.

[0041] In order to create an ideal cognitive model 160, it is preferable to create a thought reproduction system at the time of birth of the target subject, and input all subsequent biological information, behavioral information, and stimulus information. In the present mobile wireless communication device 100, 10 to approximately 10,000 parallelized thought-reproducing AIs are incorporated into one person's cognitive model 160. Each of these thought-reproducing AIs represents a part of the target person's diverse brain patterns and personality. For example, a visually dominant brain pattern, an auditory dominant brain pattern, a forgetful brain pattern, a high IQ brain pattern, or a brain pattern with strong genetic characteristics are each represented by a thought-reproducing AI. The target person is assumed to always possess multiple brain patterns, and one of these brain patterns is activated depending on the environment and situation.

[0042] The 10 to approximately 10,000 parallelized thought reproduction AIs are independent of each other. In each thought reproduction AI, the weight coefficients of each neural network are calculated based on the input 220 (input information), just as in the case of a single thought reproduction AI. As with a single thought replication AI, each thought replication AI can be updated based on continuous input 220 using backpropagation techniques. The activity level of each thought reproduction AI is evaluated based on the distance between the information (vector) about the target subject and the output vector from the thought reproduction AI. The closer the distance between the vectors, the more likely it is to be activated, and the thought reproduction AIs are sorted in order of highest activity. This distance is calculated using Euclidean distance or cosine similarity. At regular intervals, activity levels are reevaluated using new input information, and the order of the thought reproduction AIs is updated.

[0043] (B) Second program 131B (behavior prediction system, behavior prediction AI) FIG. 10 is a conceptual diagram of a behavior prediction system configured by the second program 131B. The behavior prediction system is equipped with a behavior prediction AI, which is connected to a thought reproduction AI and forms a pair with it. The behavior prediction AI predicts the thoughts and actions of a target subject for a time span T after a certain time. The behavior prediction system has the function of inputting 220 (see FIG. 3) the thought patterns or behavior patterns of the target subject output from the first program 131A (thought reproduction AI), numerically analyzing the time correlation from the input 220, and predicting the future thoughts and behaviors of the target subject. Numerical analysis methods include methods that use statistics or machine learning to handle time correlation functions, DL, which is statistical processing (AI) using huge amounts of data, and large-scale language models (LLM). Since analysis methods are rapidly evolving, it is preferable to adopt the most appropriate method at the time. For example, this behavior prediction system uses a method that uses a neural network.

[0044] The time span T is not limited to one, and multiple spans can be selected. When multiple spans are selected, the time span T can be selected arbitrarily, but typical examples include five spans: 10 seconds, 1 hour, 1 week, 1 month, and 3 months. In general, a shorter time span T is more suitable for predicting specific individual behavior, while a longer time span T is more suitable for predicting organizational trends. By applying the backpropagation method to behavior prediction AI based on continuous input information, the weight coefficients of the neural network are updated every time the time span T of future predictions is reached, thereby updating the behavior prediction system. The behavior prediction system calculates 10 to about 100 predicted behaviors that will occur after a time span T.

[0045] If there are 10 predicted behaviors, then for one target subject, a pair of thought reproduction AI and behavior prediction AI will be used to predict 10 behaviors for each of five time ranges (10 seconds, 1 hour, 1 week, 1 month, and 3 months later), resulting in a total of 1 x 5 x 10 = 50 predicted behaviors. The most appropriate predicted action is selected from these 50 predicted actions. Two indicators used to select the most appropriate predicted behavior are, for example, "Most Likely CoA (MLCoA)" and "Most Dangerous CoA (MDCoA)." One method for calculating the "most probable action (MLCoA)" is to use a softmax function in the output layer to calculate the probability of each action. One method for calculating the "Most Risky Behavior (MDCoA)" is to first define criteria for quantitatively assessing the risk of each behavior, create a database of statistical risk assessments based on past data, and then refer to this database to calculate the risk of predicted behavior. A corresponding action prediction AI is constructed for each thought reproduction AI, and the two are connected. This allows a unique prediction pattern to be generated for each thought reproduction AI.

[0046] As with a single behavioral prediction system, the behavioral prediction system is updated using the backpropagation method for each behavioral prediction AI based on continuous input information. As mentioned above, each thought reproduction AI creates 10 to 100 predicted actions for each time span T, resulting in a total of 10 x 5 x 10 = 500 predicted actions. From these 500 predicted actions, the most appropriate predicted action is selected based on the two criteria mentioned above: "Most Probable Action (MLCoA)" and "Most Risky Action (MDCoA)." From the perspective of analytical philosophy, it is said that "there are 10 possible worlds for each person 10 minutes later."

[0047] A cognitive model consisting only of thought-reproduction AI can only make inferences based on the target subject's past and present information. In contrast, a cognitive model consisting of a pair of thought-reproduction AI and behavior-prediction AI adds the element of time, further strengthening the inference results. If we interpret brain patterns as activated parts of the brain and thought patterns as emotions (or feelings), then there is the following correlation between the two. TIFF2026017599000002.tif57141

[0048] Also, when using images to trigger these thought patterns (emotions), the following are some examples of images: TIFF2026017599000003.tif78152

[0049] (C) Third Program 131C (Behavior Change System, Behavior Change AI) FIG. 11 is a conceptual diagram of a behavior modification system (behavior modification AI) configured by the third program 131C. The behavior change system is an inference system that designs intervention measures to change the behavior of a specific target subject when a target behavior (target behavior) is set for that target subject using the above-mentioned thought reproduction system and behavior prediction system, such as ``we want the target subject to perform a specific behavior (target behavior) after a specific time has passed.'' One behavior change system is built for each target individual. In other words, one behavior change AI is built for each target individual's cognitive model, and it is not parallelized like thought reproduction AI or behavior prediction AI. The input information used in building behavioral modification AI is experimental data from subjects who were subjected to various interventions used in cognitive warfare. This experimental data includes information on the stimuli that affected the subjects, as well as their behavior and biological information after receiving the stimuli.

[0050] Intervention actions include liminal and subliminal effects through mass media, social media, verbal inducements, and images and videos. Medical devices that directly affect the human body can also be used. Brain imaging devices such as functional magnetic resonance imaging (fMRI), electroencephalography (EEG), and magnetoencephalography (MEG) can be used to directly intervene in the brains of target subjects. When initially building a behavior change AI, if the target of construction is limited to a specific individual, the information obtained will be limited. Therefore, it is possible to build a behavior change AI based on a wide range of subject data, and then customize this generalized behavior change AI to be specific to the target individual. By copying the generalized behavior change AI, it is possible to build behavior change AI that is specialized for each target audience. Once the behavior change AI begins operation, the weighting coefficients of the behavior change AI specialized for the target individual will be continuously updated using new input information when intervention actions are carried out on the target individual.

[0051] As mentioned above, the behavior change AI works in collaboration with the thought reproduction AI and behavior prediction AI to design interventions to guide target individuals toward target behaviors. The behavior change AI takes as input 220 (see Figure 3) the "target subject and the behavior (target behavior) that the target subject is desired to perform and its deadline" output from the second program 131B (behavior prediction AI) and designs as output 230 an intervention plan to encourage the target subject to perform the target behavior. The behavior change AI does not need to be limited to one intervention plan, but can design multiple intervention plans depending on the situation of the target individual. Therefore, the design and simulation of intervention plans can be carried out simultaneously for 10 to 100 plans.

[0052] The output 230 of the behavior change AI indicates the intervention policy, specific actions of the intervention, and the timing of implementing the intervention. The interventions designed by the behavior change AI are placed in a simulation environment consisting of thought reproduction AI and behavior prediction AI as input information, and are then subjected to trial and verification. Through repeated trial and verification, the interventions for the target individuals are optimized. After this, an intervention action is implemented on the target subject, encouraging a change in the target subject's natural behavior to the target behavior. When the intervention action is implemented, neural information feedback occurs at all levels of abstraction in the target subject, including symbolic, quasi-symbolic, and non-symbolic information, resulting in a change in the target subject's cognition (homeostatic entrainment). In this way, by continuously implementing intervention actions, the person's cognitive function in the real world approaches the behavior (target behavior) designed by the cognitive model 160 of the behavior change induction system 300.

[0053] (Intervention Delivery System) However, implementing interventions presents many challenges, especially when it comes to targeting people, as existing methods have limitations. For example, when mass media is used to target a wide range of people (target groups), each person has different brain patterns and other biological information, so if the same method is used for all people, it is not possible to expect a uniform intervention effect. In addition, for interventions targeting a single target individual, the use of brain imaging devices is effective, but this is not practical because it requires the target individual's consent to use the device and requires enormous computer resources to analyze large amounts of data.

[0054] Therefore, in order to solve these problems, the behavior change induction system 300 according to this embodiment employs an intervention delivery system that makes it possible to respond to a large number of target subjects. Figure 12 is a schematic diagram of the intervention delivery system. In this intervention delivery system, interventions are separated into highly abstract intervention policy signals and specific intervention actions. By separating intervention measures into intervention policy signals and intervention actions, it is also possible to separate the AI ​​for behavior change. For example, the function of designing intervention policy signals can be performed by a behavior change AI located on a cloud server, while the generation and implementation of intervention actions can be performed by an intervention implementation AI located in an application on a device (such as a mobile phone) owned by the target subject.

[0055] Specifically, as shown in Figure 12, the intervention delivery system consists of a behavior change AI 171 on a cloud server 170 that generates an intervention policy signal, and an intervention implementation AI 174 on an application in a portable wireless receiving device (e.g., a smartphone) 173 carried by a target subject 172. The behavior change AI 171 (third program 131C) on the cloud server 170 works in cooperation with the first program 131A (thought reproduction AI) and the second program 131B (behavior prediction AI) to design an intervention policy for behavior change of the target subject 172. The intervention policy is encrypted and compressed to become a lightweight intervention policy signal of about 8 bytes, such as MIDI data. This intervention policy signal contains information with a certain degree of abstraction, and is set so that the data volume can be small. The intervention strategy signal is transmitted to a portable wireless communication device 173 carried by a target subject 172 via an existing information and communication infrastructure (the Internet).

[0056] When the intervention implementation AI 174 in the portable wireless communication device 173 receives the intervention plan signal from the behavior change AI 171, it formulates an intervention action customized for each target subject based on the intervention plan signal and puts it into action. Examples of such interventions include altering the text of chat messages, issuing fake earthquake alerts or other disaster warnings, and displaying subliminal images on the display of the portable wireless communication device 173. Furthermore, it is also possible to obtain information (operation history, browsing history, center information, etc.) of the target subject 172 that can be obtained through the portable wireless communication device 173 via an application (intervention implementation AI 174). For example, if the above-mentioned intervention delivery system is used in the marketing field, the intervention can be divided into an intervention policy signal, such as "getting potential customers to view the campaign page for your company's product within one hour," and an intervention action, such as "displaying a push notification on the display of the potential customer's portable wireless communication device (such as a smartphone)," to achieve that objective.

[0057] Furthermore, when the intervention delivery system is applied to the military domain, intervention measures can be divided into intervention policy signals, such as "making the enemy leader abandon military invasion within six months," and intervention actions, such as "making an official announcement at an international organization to create international public opinion that military invasion is unjustified," in order to achieve that goal. Figures 13, 14, and 15 are examples of inputs and outputs (intervention policy signals and intervention actions) in the security domain, marketing domain, and entertainment domain, respectively. [Example]

[0058] Below, an example will be given in which the behavior modification induction system 300 according to this embodiment is more specifically applied.

[0059] (First Example) In a first embodiment, the behavior change induction system 300 is configured as a cognitive warfare defense system. Target subjects in a cognitive warfare defense system are, for example, the leaders of an enemy country, military personnel of an enemy country, citizens of an enemy country, and nationals of another country (or countries) who are the subject of international public opinion. When computer resources are sufficient, it is preferable to build a cognitive model for the entire population, but when computer resources are limited, the target population is limited to one or a few people. For example, the leader of an enemy country is limited to the enemy's prime minister and several close aides who have influence on the prime minister's decisions, and the military personnel of an enemy country are limited to the commander of the enemy's military and his close aides who have influence on the commander's decisions.Furthermore, for the citizens of an enemy country and the nationals of other countries who are the main players in international public opinion, a cognitive model is constructed as the collective cognition of people belonging to a certain age group, residents of a specific region, or nationals of a specific country.

[0060] For example, if the target action against an enemy country is to prevent war, examples of intervention intention signals would be to make the enemy's leaders feel that starting a war would mean losing their position, to make the enemy's military personnel feel that they have no chance of winning in terms of military strength and logistics, and to make the enemy's citizens feel that war will make their lives more difficult. Another example is encouraging the people of other countries, who are the main actors in international public opinion, to speak out in condemnation of the enemy country's declaration of war.

[0061] (Second Example) In the second embodiment, the behavioral change induction system 300 is configured as a marketing system. The goal of the marketing system is to understand consumer behavior and expand its influence based on that understanding. The behavior change induction system 300 reproduces the consumer's decision-making process, preferences, needs, and behavioral patterns, making it possible to design and implement a personalized marketing strategy tailored to the target audience (potential customers). In this marketing strategy, for example, the following steps are implemented: First, a cognitive model 160 (see FIG. 5) of each consumer (target audience) is constructed. Specifically, an individual cognitive model is constructed for each consumer based on biometric information such as the consumer's brain patterns, as well as information related to purchasing behavior such as the consumer's past purchasing history, online behavioral trajectory, and reactions and opinions expressed on social media.

[0062] This cognitive model 160 makes it possible to infer what cognitive biases consumers have toward specific products and services, or what factors promote or inhibit their desire to purchase. Second, a personalized marketing strategy is designed. Specifically, the constructed cognitive model 160 is used to plan marketing messages and promotional activities optimized for the current needs and future desires of consumers suggested by the behavioral prediction AI (second program 131B). This allows the deep psychology and latent needs of consumers to be reproduced in the marketing field, and an effective approach corresponding to them is created, which can directly lead to an improvement in the company's brand value and an increase in sales.

[0063] (Third Example) In the third embodiment, the behavior change induction system 300 is configured as an entertainment system. By applying the behavior change induction system 300 to the entertainment field, it is possible to provide a personalized experience tailored to the individual tastes and behavior patterns of a wide range of target audiences, such as movie watchers and game and music enthusiasts. In this entertainment system, for example, the following steps are performed. First, a cognitive model 160 of each viewer or other target audience is constructed. Specifically, an individual cognitive model 160 for each target subject is constructed based on biometric information such as brain patterns, heart rate, skin conductance, electroencephalograms, and pupil response when the audience watches a film, as well as the target subject's past movie viewing history, game play records, and music streaming information. This cognitive model 160 makes it possible to infer which genres and themes each target subject is particularly interested in, what types of stories and gameplay they prefer, and what moves them.

[0064] Second, personalized content is recommended. The cognitive model developed is used to recommend movies, games, music, etc. that are best suited to each target audience, and can even recommend content at a detailed level, such as the visual presentation, game story progression, and musical tone. Third, the interactive content experience is customized based on the viewer's cognitive model. For example, movie scenes that viewers have had strong emotional reactions to in the past could be analyzed to recommend content that evokes similar emotions, or in-game events could be tailored to suit the player's preferences. Fourth, the user experience is continuously optimized. By incorporating feedback from the target audience and new behavioral data of the target audience into the cognitive model, the target audience's experience is continuously optimized, allowing it to quickly respond to users' changing and emerging preferences, always providing the best entertainment experience.

[0065] In this way, by applying the behavior change induction system 300 to the entertainment field, the entertainment industry can provide more personalized and immersive experiences, significantly improving the satisfaction and loyalty of viewers and players. Furthermore, this approach is expected to promote the discovery of new content and expand entertainment consumption. To facilitate understanding of the behavioral modification induction system 300 according to this embodiment, a specific example will be given below.

[0066] (Example 1) 1. Target Audience Individual A is the target audience. 2. Users of this behavior change guidance system 300 It is assumed that a friend or family member of person A uses this behavior change induction system 300 to try to change person A's behavior. 3. Training data The following input data (stimulus) and output data (response) are used as training data 200 (see FIG. 3). The input data (stimuli) given to individual A are as follows: (A) Showing scary images (movies, TV, etc.) (B) Telling scary stories (such as reading thriller novels) The responses (output data) shown by Individual A to these input data (stimuli) were as follows: (1) Your heart beats faster (2) Breaking into a cold sweat (3) Trying to escape

[0067] These teacher data 200, along with other biometric information of individual A, are sent to the evaluation model generation program 135B by the teacher data input program 135A of the first program 131A, and the evaluation model generation program 135B generates an evaluation model 210 through machine learning (see step S120 in Figure 4). 4. Behavioral prediction using behavioral prediction AI After the evaluation model 210 has undergone learning, the second program 131B (behavior prediction AI) predicts how individual A will behave in response to real-world stimuli. When individual A encounters the following stimuli (a) to (c), the second program 131B (behavior prediction AI) makes the following behavior prediction. (a) Person A hears a ghost story

[0068] The behavior prediction AI predicts that individual A will behave in response to this stimulus as either "no reaction" or "his heartbeat will speed up slightly." (b) While walking down a street at night, Individual A mistook something for a ghost. The behavior prediction AI predicts that Individual A will "fall into a state of mental paralysis (go blank)" in response to this stimulus. (c) While walking in the mountains, a snake suddenly appeared in front of me. In response to this stimulus, the behavior prediction AI predicts that Individual A will either "try to escape but his legs will stop moving" or "his legs will give way." 5. Behavioral change actions using behavioral change AI Based on the behavior prediction of individual A output by the behavior prediction AI, third program 131C (behavior change system) formulates behavior change actions to change the behavior of individual A.

[0069] There are two ways to change behavior: (1) Behavioral change that weakens the emotion (affect) caused by a stimulus (attenuating behavioral change) (2) Behavioral change that replaces the emotion (emotion) caused by a stimulus with another emotion (emotion) (substitutional behavioral change) The behavior change actions formulated by the third program 131C (Behavior Change System) are either or both of these. (5.1) For example, suppose that person A decides to go out in the middle of the night. In this case, the weakening behavior change action formulated by the behavior change AI (the behavior change action in response to "Individual A mistaking something for a ghost") would be, for example, as follows:

[0070] (A) Have individual A wear a helmet equipped with a powerful light, put the helmet on, and walk while illuminating the area ahead. If the area ahead is bright, the chances of "Individual A mistaking something for a ghost" are significantly reduced. The third program 131C (behavior change system) notifies the user of the behavior change guidance system 300 (friends or family of individual A) of this behavior change action, for example, via email, or displays it on a display. This allows the user of the behavior change guidance system 300 to purchase a helmet with a powerful light and give it to individual A. Alternatively, if there is time, the third program 131C (behavior change system) can automatically order a lighted helmet from a retail website (e.g., an online store such as Amazon (registered trademark)) and have it delivered by express to the user of this behavior change induction system 300 or directly to individual A.

[0071] (B) A conversation continues between the automated voice and individual A via the portable wireless communication device of individual A. That is, the third program 131C (behavior modification system) performs a voice analysis of the words uttered by individual A, and via the automated voice system, creates corresponding words and outputs them from individual A's portable wireless communication device. This forces person A to focus on the conversation and not have time to mistake something for a ghost. Alternatively, an icon can appear on the display of person A's mobile phone, and person A can have a conversation with the icon. (C) The third program 131C (behavioral modification system) utilizes the map function of the portable wireless communication device of the individual A to search for a well-lit route to the destination, and guides the individual A to the destination using the portable wireless communication device of the individual A. Avoiding dark roads reduces the chance of misreading. Alternatively, instead of a well-lit route, a route with lots of people can be adopted. The presence of people creates a sense of security, reducing the chance of mistakes that often result from anxiety.

[0072] Next, the replacement behavior change actions formulated by the behavior change AI might look something like this: (D) Beautiful night cityscapes and soothing music are played on Individual A's portable wireless communication device to elicit positive emotions. This not only alleviates individual A's fear, but also brings out positive emotions in individual A by allowing them to enjoy beautiful scenery and pleasant music. (E) Encouraging video calls with friends and family via Individual A's mobile wireless communication device, providing Individual A with a sense of security and a sense of human connection and solidarity. In a situation that would normally frighten Individual A, this helps them feel more secure by making them aware of their connections with friends and family.

[0073] (F) Displaying photos and videos of past happy memories on the display of Individual A's portable wireless communication device elicits positive emotions. In a situation where individual A would normally feel fear, by replaying a pleasant memory from the past, individual A's fear is overwritten with positive emotions. (5.2) For example, suppose that individual A decides to go hiking. In this case, the weakening behavior change action (the behavior change action in response to "a snake suddenly appearing in front of you") formulated by the behavior change AI would be as follows: (A) Have individual A wear sturdy (highly impact-resistant) boots. Since the majority of snake bites occur on the feet, this is to give Individual A the peace of mind that as long as he protects his feet, he will be fine.

[0074] The third program 131C (behavior change system) notifies the user of the behavior change induction system 300 (a friend or family member of individual A) of this behavior change action, for example, via email, or displays it on a display. This allows the user of the behavior change induction system 300 to purchase the boots and have individual A wear them. Alternatively, if there is time, the third program 131C (behavior change system) can automatically order the boots from a retail website (e.g., an online store such as Amazon (registered trademark)) and have them delivered by express to the user of this behavior change induction system 300 or to individual A.

[0075] (B) Have individual A wear clothing that prevents heat spread. Snakes use their tongues to sense heat and identify prey, so this is to prevent heat from being released to the outside. The third program 131C (behavior change system) notifies the user of the behavior change induction system 300 (friends or family of individual A) of this behavior change action, for example, via email, or displays it on a display. This allows the user of the behavior change induction system 300 to purchase heat diffusion prevention clothing and have individual A wear it. Alternatively, as in case (A), it is also possible to place an automatic order on a product sales website. (C) The third program 131C (behavioral modification system) sends a visual or audio suggestion to individual A that "snakes are not scary" via the display 142 or speaker 145 of individual A's portable wireless communication device. Even if a human encounters a snake, the snake will most likely run away unless the human attacks, so this is to prevent Individual A from feeling more fear than necessary.

[0076] Next, the replacement behavior change actions formulated by the behavior change AI might look something like this: (D) Positive information about the beautiful scenery and flora and fauna of the mountains is displayed on the display of Individual A's portable wireless communication device, allowing Individual A to enjoy the scenery while walking. By overriding the fear of snakes with information about beautiful scenery and interesting flora and fauna, the fear of snakes is alleviated and hiking becomes more enjoyable. For example, a user of the behavioral change induction system 300 can send instructions to the third program 131C (behavior change system) of individual A's portable wireless communication device via individual A's portable wireless communication device, and display these instructions.

[0077] (E) An audio guide emphasizing the health and relaxation benefits of hiking is played from Individual A's portable wireless communication device. Override fear of snakes with positive health and relaxation effects. For example, a user of the behavioral change induction system 300 can send instructions to the third program 131C (behavioral change system) of individual A's portable wireless communication device via individual A's portable wireless communication device and play these audio instructions. (F) Displaying photos or videos on the display of Individual A's portable wireless communication device that evoke memories of past enjoyable hikes or nature explorations. In situations that would normally frighten Individual A, replaying memories of past enjoyable hikes and nature explorations overwrites the fear of snakes with positive emotions. For example, a user of the behavioral change induction system 300 can send instructions to the third program 131C (behavior change system) of individual A's portable wireless communication device via individual A's portable wireless communication device, and display these instructions.

[0078] (Example 2) 1. Target Audience The target person is President P of country R. 2. Users of this behavior change guidance system 300 The government of country F, which is in a conflict with country R, ​​uses this behavior change induction system 300 to try to change the behavior of country R's president P. 3. Training data The following input data (stimulus) and output data (response) are used as training data 200 (see FIG. 3).

[0079] The input data (stimuli) given to the target subject are as follows: (A) The domestic economy is sluggish (the price of crude oil, which is R's source of foreign currency, has fallen, and foreign currency is not coming into R's country). (B) A new large oil field is discovered in another country, or a shale revolution occurs in another country (this will further lower crude oil prices, further hurting the economy of Country R, ​​which is dependent on crude oil). The responses (output data) shown by the target subjects to these input data (stimuli) were as follows: (1) Engaging in military activities abroad (such as annexing the territory of other countries or intervening militarily in conflicts in other countries) to divert public discontent caused by the sluggish domestic economy. (2) Strengthening the domestic intelligence network (exposing dissatisfied elements within the country) (3) Strengthening military power

[0080] 4. Behavioral prediction using behavioral prediction AI After the evaluation model 210 has undergone learning, the second program 131B (behavior prediction AI) predicts how the target subject will behave in response to real-world stimuli. When the target subject encounters the following stimulus, the second program 131B (behavior prediction AI) makes the following behavior prediction. (a) Country S, which had maintained neutrality toward both country R and its enemy country F, abandons its neutral policy and joins a military alliance N (e.g., NATO) to which country F belongs. In response to this stimulus, the behavior prediction AI predicts that the target person will behave as follows: "In order to prevent the further expansion of military alliance N, invade neighboring country U, which has not yet joined military alliance N, and annex neighboring country U into country R."

[0081] 5. Behavioral change actions using behavioral change AI In response to the predicted behavior of the target individual (invasion of neighboring country U) output by the behavior prediction AI, the third program 131C (behavior change system) formulates behavior change actions to change the behavior of the target individual. As mentioned above, the third program 131C (Behavior Change System) develops both attenuating and replacement behavior change actions. Examples of debilitating behavioral change actions formulated by behavioral change AI include the following: (A) Request the leaders of each country to strengthen military support for Country U. Specifically, the third program 131C (behavioral change system) encourages Country F, which is a user of the behavioral change induction system 300, to issue a statement or send an email to the leaders of other countries. (B) Encourage the media to raise international public opinion. For example, the third program 131C (behavioral change system) allows country F, which is a user of this behavioral change induction system 300, to provide information about the movements and intentions of country R to media outlets and other media in each country, encouraging them to publish articles of protest and criticism against country R's planned invasion of country U.

[0082] (C) Announce economic sanctions (a boycott of oil from Country R). For example, the third program 131C (behavioral change system) urges Country F, which is a user of the behavioral change induction system 300, to issue a statement announcing economic sanctions or to issue a notice to the leaders of other countries requesting their participation in the economic sanctions. (D) Countries hostile to R will issue a statement emphasizing their support for U and fostering a sense of international solidarity. This will make the target feel internationally isolated, fear diplomatic failure, and abandon their invasion of Country U. For example, the third program 131C (behavioral change system) urges Country F, a user of the behavioral change induction system 300, to notify the leaders of other countries of the holding of an international conference to issue a statement, or to request that the media outlets of other countries publish articles calling for the holding of an international conference.

[0083] (E) Propose an international economic cooperation program to assist Country R's economic recovery. If Country R's economy recovers, the need to invade Country U to turn its people's attention abroad will gradually decrease. For example, the third program 131C (behavioral change system) urges Country F, a user of the behavioral change induction system 300, to present such a policy and issue a notice calling for cooperation from other countries. Furthermore, it urges Country F to request that media outlets in each country publish articles highlighting the need for such a program. (F) Conduct information warfare to suppress anti-government activity within the R and emphasize political stability within the country. If the domestic political situation stabilizes, there will be less need to invade Country U to turn the nation's attention abroad. For example, the third program 131C (behavioral change system) presents such a policy to country F, which is a user of the behavioral change induction system 300, and encourages it to issue a notice calling for cooperation from other countries (allies).

[0084] Next, the replacement behavior change actions formulated by the behavior change AI might look something like this: (G) Providing international conferences and negotiations that highlight diplomatic successes for target audiences and elicit a sense of euphoria from diplomatic successes. By emphasizing cooperation and support for Country R from the international community and making the target feel like they have achieved diplomatic success, the motivation for invasion is reduced. For example, the third program 131C (behavioral change system) urges Country F, which is a user of the behavioral change induction system 300, to notify the leaders of each country of the holding of an international conference, or to request the media to publish articles in each country's news organizations calling for the holding of an international conference. (H) Propose economic policies and infrastructure projects that will lead to an increase in the approval rating of the target person (President P) in R, and elicit a sense of happiness from the target person due to the increase in approval rating. By emphasizing domestic economic stability and infrastructure projects and gaining public support, they can make invasion seem less necessary. For example, the third program 131C (behavioral change system) presents such policies and projects to Country F, which is a user of the behavioral change induction system 300, and encourages Country F to notify Country R of the policies and projects.

[0085] (I) Promote the holding of cultural and sporting events that will earn the target audience national and international acclaim, and build a positive international image of the target audience. Through international sporting events and cultural exchanges, the target audience will be given a positive international image and admiration, making them feel less of a need for invasion. For example, the third program 131C (behavioral change system) proposes to Country F, a user of the behavioral change induction system 300, plans for such a sporting event and encourages the country to notify its leaders of their participation in the sporting event, or to request the media of each country to publish articles advocating the need to hold the sporting event.

[0086] (J) Providing new diplomatic initiatives to create a sense of international success for target audiences. By making their targets feel internationally supported and highlighting their diplomatic successes, they can ease their fears of being overthrown and discourage them from invading Country U. For example, the third program 131C (behavioral change system) urges country F, which is a user of this behavioral change induction system 300, to notify the leaders of other countries of the holding of an international conference to provide country R with a diplomatic initiative, or to request the media of each country to publish an article calling for the holding of such an international conference.

[0087] (K)Provide success stories and positive data that highlight the economic growth and domestic stability of the country, creating a sense of economic stability. By easing economic fears and overwriting them with data that suggests domestic stability and growth, the plan will discourage people from invading Country U. For example, the third program 131C (behavioral change system) presents such examples and data to Country F, which is a user of the behavioral change induction system 300, and has Country F convey them to the target subjects. (L) Provide target audiences with successful examples of cultural exchanges and sporting events that are internationally acclaimed, and highlight a positive international image. Through international success stories, the goal is to elicit positive emotions from the target audience and alleviate their fears, thereby discouraging them from invading Country U. For example, the third program 131C (behavioral change system) presents emails and videos showing such cases to Country F, a user of this behavioral change induction system 300, and Country F then sends those emails and videos to news organizations in each country.

[0088] (Example 3) 1. Target Audience The target audience is heavy viewers of videos via mobile wireless communication devices (smartphones) (who rarely go to the cinema). 2. Users of this behavior change guidance system 300 It is assumed that a friend or family member of the target subject uses the behavior change induction system 300 to try to change the behavior of the target subject. 3. Training data The following input data (stimulus) and output data (response) are used as training data 200 (see FIG. 3). The input data (stimuli) given to the target subject are as follows: (A) Show your favorite short video clip (B) Have them watch the video while listening to relaxing music. (C) Providing easy-to-enjoy video content The responses (output data) shown by the target subjects to these input data (stimuli) were as follows: (1) Smile (2) Take a relaxed posture (3) Feeling comfortable

[0089] 4. Behavioral prediction using behavioral prediction AI After the evaluation model 210 has undergone learning, the second program 131B (behavior prediction AI) predicts how the target subject will behave in response to real-world stimuli. When the target subject encounters the following stimulus, the second program 131B (behavior prediction AI) makes the following behavior prediction. (a) Show your target audience an exciting action movie trailer. In response to this stimulus, the behavior prediction AI predicts that the target person will "show an excited expression" or "lean forward." (b) A friend plans to go to the cinema with the target audience. In response to this stimulus, the behavior prediction AI predicts that the target person will actively say, "I want to go to the cinema." (c) Showing a target audience information about special events at the cinema. In response to this stimulus, the behavior prediction AI predicts that the target person's behavior will be "decide to go to the cinema" or "start preparing to go to the cinema."

[0090] 5. Behavioral change actions using behavioral change AI Based on the behavior prediction of the target subject output by the behavior prediction AI, the third program 131C (behavior change system) formulates behavior change actions to change the behavior of the target subject. As mentioned above, the third program 131C (Behavior Change System) develops both attenuating and replacement behavior change actions. (5.1) For example, suppose a target subject wants to laze around at home watching videos on their mobile wireless communication device (smartphone). In this case, the weakening behavior change actions formulated by the behavior change AI would be, for example, as follows:

[0091] (A) Limit video viewing time on portable wireless communication devices. For example, the third program 131C (behavioral change system) sends a reminder to the user of the behavioral change induction system 300 to limit the target subject's video viewing time, and displays a message to that effect on the display screen of the target subject's portable wireless communication device, or periodically plays an automated voice message via the target subject's portable wireless communication device. Alternatively, the user of the behavioral change induction system 300 can remotely control the portable wireless communication device of the target subject and forcibly limit the video viewing time of the target subject. (B) Inform people about the health effects of prolonged viewing. For example, the third program 131C (behavior change system) may notify the user of the behavior change induction system 300 of this fact, and the user may directly notify the target subject. Alternatively, this fact may be displayed on the display screen of the target subject's portable wireless communication device, or an automated voice message may be played periodically.

[0092] (C) Send a message encouraging a break after a certain period of video viewing. For example, the third program 131C (behavior change system) may notify the user of the behavior change induction system 300 of this fact, so that the user can urge the target subject to take a break. Alternatively, the third program 131C may display a message to that effect on the display screen of the target subject's portable wireless communication device, or may periodically play an automated voice message. Next, the replacement behavior change actions formulated by the behavior change AI might look something like this: (D) By showing trailers and behind-the-scenes footage of an exciting action movie, the target audience can anticipate that the movie will be even more exciting when they see it in the theater. For example, third program 131C (behavior modification system) may stream such images to the displays of target subjects' portable wireless communication devices, replacing the experience of watching a video at home with the excitement of a movie theater.

[0093] (E) Notify you of special events and exclusive screenings at movie theaters and encourage you to plan a trip to the theater with friends. For example, the third program 131C (behavior change system) notifies the user of the behavior change induction system 300 of this fact, and provides information from the user to the target person. Alternatively, the third program 131C (Behavior Change System) may send emails to the target audience or inform them of this information via displays or speakers, thereby emphasizing the special cinema experience and encouraging them to make plans with friends, thereby eliciting positive behavior. (F) Show your target audience photos and videos that evoke fond memories of the cinema, making them want to go back to the cinema. For example, the third program 131C (behavior modification system) displays such photos and videos on the display of the target subject's portable wireless communication device, thereby reminding the target subject of past pleasant experiences at the movie theater and activating their behavior.

[0094] (5.2) For example, suppose a target audience decides to go to a movie theater. In this case, the weakening behavior change actions formulated by the behavior change AI would be, for example, as follows: (A) Provide an application that allows users to easily search for and reserve transportation to movie theaters. For example, the third program 131C (behavioral modification system) automatically downloads such an application to the target subject's portable wireless communication device without any instruction from the target subject or the user, thereby reducing the hassle of checking travel routes and thereby lowering the barrier for the target subject to go to the cinema. (B) Provide discount coupons and special offers at movie theaters to emphasize the benefits of going to the movies. For example, the third program 131C (behavior modification system) displays discount coupons and special offer information for movie theaters on the display of the portable wireless communication device of the target subject, making the target subject feel that they will get a good deal by going to the movie theater, and encouraging the target subject to take action.

[0095] (C) Providing real-time information on crowding and screening schedules at movie theaters so you can go at the best time. For example, the third program 131C (behavior modification system) displays the information on the display of the target subject's portable wireless communication device, making the target subject expect a smooth movie theater experience and encouraging the target subject to take action. Next, the replacement behavior change actions formulated by the behavior change AI might look something like this: (D) Highlight special events and interactive experiences at the cinema, suggesting that there are things you can only experience at the cinema. For example, the third program 131C (behavior modification system) displays the information on the display of a portable wireless communication device, and replaces watching a video at home with a special movie theater experience.

[0096] (E) Suggest making plans to enjoy a movie with friends and family, emphasizing social connections. For example, the third program 131C (behavior modification system) displays such suggestions on the display of the target subject's portable wireless communication device, encouraging the target subject to view the movie theater experience as a social event and to become more active. (F) Provide video messages about past enjoyable experiences at movie theaters and new experiences people can look forward to at movie theaters, to motivate people to actively go to the movies. For example, third program 131C (behavior change system) displays such video messages on the display of the target subject's portable wireless communication device to highlight the target subject's positive experience at the movie theater and encourage the target subject to be more active.

[0097] (Example 4) 1. Target Audience The target person is President P of country R. 2. Users of this behavior change guidance system 300 Assume that the Japanese government uses this behavior change induction system 300 to attempt to change the behavior of President P of country R. 3. Training data The following input data (stimulus) and output data (response) are used as training data 200 (see FIG. 3). The input data (stimuli) given to the target subject are as follows: (A) Show documents and records showing that R's neighboring country, U, has traditionally been R's territory. (B) Country R learns that the residents of a certain area of ​​neighboring country U are originally from country R, ​​and that these residents are being mistreated by the government of country U.

[0098] The responses (output data) shown by the target subjects to these input data (stimuli) were as follows: (1) They decided that they had to help the residents of a certain area of ​​neighboring country U, and decided to invade neighboring country U. 4. Behavioral prediction using behavioral prediction AI After the evaluation model 210 has undergone learning, the second program 131B (behavior prediction AI) predicts how the target subject will behave in response to real-world stimuli. When the target subject encounters the following stimulus, the second program 131B (behavior prediction AI) makes the following behavior prediction. (a) I have come across documents and records that show that the Ainu people, who currently live in Hokkaido, Japan, were originally residents of Country R and were the indigenous people of Japan. (b) The Ainu people are being mistreated by the Japanese government. In response to these stimuli (inputs), the behavior prediction AI predicts that the target person's behavior will be "invade Hokkaido, annex it into Country R, ​​and save the Ainu people."

[0099] 5. Behavioral change actions using behavioral change AI In response to the predicted behavior of the target individual (invasion of Hokkaido) output by the behavior prediction AI, the third program 131C (behavior change system) formulates behavior change actions to change the behavior of the target individual. As mentioned above, the third program 131C (Behavior Change System) develops both attenuating and replacement behavior change actions. Examples of debilitating behavioral change actions formulated by behavioral change AI include the following: (A) Send visible signals (issuing a statement to the international community, raising the issue at the UN, and strengthening the Self-Defense Forces' deployment in Hokkaido). For example, the third program 131C (behavioral change system) makes these proposals to the Japanese government, which is the user of this behavioral change induction system 300. (B) Encourage the media to raise international public opinion. For example, the third program 131C (behavioral change system) urges the Japanese government, which is a user of the behavioral change induction system 300, to provide the necessary information to the media and request that they publish an article.

[0100] (C) Announce economic sanctions (a boycott of Country R's oil) through public statements and the media. For example, the third program 131C (behavioral change system) makes these proposals to the Japanese government, which is the user of this behavioral change induction system 300. Next, the replacement behavior change actions formulated by the behavior change AI might look something like this: (D) By supporting Country U, which is currently at war with Country R, ​​we will strengthen Country U's offensive and force Country R to abandon its invasion of Hokkaido. (E) Offer beneficial economic assistance to Country R, ​​creating a detente between your country and Country R. In either case, for example, the third program 131C (behavior change system) formulates such a plan and proposes it to the Japanese government, which is the user of this behavior change induction system 300. [Explanation of symbols]

[0101] 100 Portable wireless communication device 110 Communications Department 120 control section 130 External memory (hard disk) 131 Application Section 132 Data storage unit 140 Input / output section 150 antenna

Claims

1. A behavior change induction system that designs intervention measures to induce behavior change in a target subject, a storage means for storing biometric information including a target subject's response to past external stimuli; a thought pattern reproducing means for reproducing the thought pattern or behavior pattern of the target person using a prediction model that has been machine-learned using the biometric information as training data; a reaction estimation means for receiving a current external stimulus or a predicted future external stimulus for the target subject as an input, and outputting a predicted reaction of the target subject to the external stimuli based on the thought pattern or behavior pattern of the target subject reproduced by the thought pattern reproduction means; an intervention design means for designing an intervention to guide the behavior of the target subject toward a target behavior in accordance with the predicted response of the target subject output by the response estimation means; A behavioral change induction system that includes:

2. 2. The behavioral change induction system according to claim 1, wherein the biological information includes brain pattern information of the target subject.

3. The behavioral change induction system described in claim 1, characterized in that the intervention design means designs at least one of an intervention to guide the target subject to weaken the emotion caused by the external stimulus and an intervention to guide the target subject to replace the emotion caused by the external stimulus with another emotion.

4. A portable wireless communication device incorporating the behavioral change induction system according to any one of claims 1 to 3.

5. A behavior change promotion system that designs and implements intervention measures to induce behavior change in target subjects, The behavioral change promotion system includes a behavioral change induction system and a portable wireless communication device owned by the target person, The behavioral modification induction system comprises: a storage means for storing biometric information including a target subject's response to past external stimuli; a thought pattern reproducing means for reproducing the thought pattern or behavior pattern of the target person using a prediction model that has been machine-learned using the biometric information as training data; a reaction estimation means for receiving a current external stimulus or a predicted future external stimulus for the target subject as an input, and outputting a predicted reaction of the target subject to the external stimuli based on the thought pattern or behavior pattern of the target subject reproduced by the thought pattern reproduction means; an intervention plan design means for designing an intervention policy for an intervention plan to guide the behavior of the target subject to a target behavior in accordance with the predicted reaction of the target subject output by the reaction estimation means; Equipped with the portable wireless communication device has an application embedded therein that designs an intervention plan based on the intervention policy; the behavior change guidance system transmits the intervention plan to the portable wireless communication device of the target subject; A behavior change promotion system in which the application of the portable wireless communication device that receives the intervention policy formulates an intervention action customized for each of the target individuals based on the intervention policy.

6. 6. The behavioral change promotion system according to claim 5, wherein the biological information includes brain pattern information of the target subject.

7. The behavioral change promotion system described in claim 5, characterized in that the intervention plan design means designs at least one of an intervention plan for guiding the target subject to weaken the emotion caused by the external stimulus and an intervention plan for guiding the target subject to replace the emotion caused by the external stimulus with another emotion.

8. A behavior change induction method for designing an intervention plan for inducing behavior change in a target subject, comprising: a first step of storing biometric information including responses of the target subject to past external stimuli; a second process of reproducing the thought patterns or behavior patterns of the target subject using a prediction model trained by machine learning using the biometric information as training data; a third process that receives current or predicted future external stimuli for the target subject as input and outputs a predicted response of the target subject to the external stimuli based on the thought patterns or behavior patterns of the target subject reproduced in the second process; a fourth step of designing an intervention plan to guide the behavior of the target subject toward a target behavior in accordance with the predicted response of the target subject output in the third step; A behavioral change induction method comprising:

9. The method for inducing behavioral change described in claim 8, characterized in that the fourth step involves designing at least one of an intervention plan for inducing the target subject to weaken the emotion caused by the external stimulus and an intervention plan for inducing the target subject to replace the emotion caused by the external stimulus with another emotion.

10. A behavior change induction method for designing an intervention plan for inducing behavior change in a target subject, comprising: a first step of storing biometric information including responses of the target subject to past external stimuli; a second process of reproducing the thought patterns or behavior patterns of the target subject using a prediction model trained by machine learning using the biometric information as training data; a third process that receives current or predicted future external stimuli for the target subject as input and outputs a predicted response of the target subject to the external stimuli based on the thought patterns or behavior patterns of the target subject reproduced in the second process; a fourth step of designing an intervention policy for an intervention plan to guide the behavior of the target subject toward a target behavior, in accordance with the predicted response of the target subject output in the third step; a fifth step in which an application installed in a portable wireless communication device owned by the target subject formulates an intervention action customized for each of the target subjects based on the intervention policy; A behavioral change induction method comprising:

11. The method for inducing behavioral change described in claim 10, characterized in that the fourth step involves designing at least one of an intervention plan for inducing the target subject to weaken the emotion caused by the external stimulus and an intervention plan for inducing the target subject to replace the emotion caused by the external stimulus with another emotion.

12. A program for causing a computer to execute the behavioral modification induction method according to claim 8 or 9.

Images