Response assistance device, response assistance method, and response assistance program

Abstract

This response assistance device estimates the degree of self-disclosure and the magnitude of stress of a consulter on the basis of the utterance content and nonverbal information of the consulter in counseling. Then, the response assistance device calculates, on the basis of the estimated magnitude of the stress of the consulter, a first value indicating the degree to which a counselor should listen to or accept the utterance of the consulter. In addition, the response assistance device calculates, on the basis of the estimated degree of self-disclosure of the consulter, a second value indicating the degree to which the counselor should sympathize with the utterance of the consulter. Then, the response assistance device generates and outputs an instruction to generate a response, to the consulter, indicating listening or acceptance at a degree corresponding to the first value, and an instruction to generate a response, to the consulter, indicating sympathy at a degree corresponding to the second value.

Description

Response support device, response support method, and response support program

[0001] The present invention relates to a response support device, a response support method, and a response support program in counseling.

[0002] In counseling, it is important for the counselor to listen and accept the client (for example, saying things like, "That must have been difficult," or "I accept you as you are"). However, listening and acceptance alone can result in a superficial response to the client, which can be counterproductive. Therefore, the counselor needs to provide feedback (i.e., empathy) that they understand the client at the appropriate time. In other words, in counseling, it is necessary to listen, accept, and empathize appropriately according to the client's characteristics and condition.

[0003] There are two types of empathy in counseling: counselors demonstrate emotional empathy to an appropriate extent, while focusing on cognitive empathy in order to accurately understand the client's condition.

[0004] Emotional empathy involves changing one's own emotions and understanding the emotions of the other person by synchronizing with them. In counseling, it is important to show a moderate amount of emotional empathy in order to build rapport with the client. However, false emotional empathy or excessive emotional empathy can be counterproductive. Cognitive empathy is understanding the client's condition on a cognitive basis without changing one's own emotions. In counseling, it is important to use cognitive empathy accurately and carefully in order to identify the client's main complaint.

[0005] However, some people feel that consulting a counselor is a psychologically and physically daunting task. For such people, various counseling AIs and robots that respond in a human-like manner have been proposed (see Non-Patent Documents 1 and 2).

[0006] Yoon Kyung Lee et al., Developing Social Robots with Empathetic Non-Verbal Cues Using Large Language Models, [online], [Retrieved June 30, 2024], Internet <URL: https: / / arxiv.org / abs / 2308.16529v1> Self-counseling using ChatGPT. We have released "mimo AI," which suggests your own cognitive biases. , [online], [Retrieved June 30, 2024], Internet <URL: https: / / prtimes.jp / main / html / rd / p / 000000004.000070803.html>

[0007] However, the above counseling AI and robots are unable to appropriately adjust the degree of listening, acceptance, and empathy to suit the client's condition, and therefore are unable to provide appropriate empathic intervention to the client.

[0008] Therefore, an object of the present invention is to solve the above-mentioned problems and to provide a means for appropriately adjusting the degree of listening, acceptance, and empathy when responding to a counselor in accordance with the state of the counselor.

[0009] In order to solve the above-mentioned problems, the present invention is characterized by comprising a state estimation unit that estimates a level of self-disclosure and a level of stress of a client based on the content of the client's utterances and non-verbal information during counseling; a first calculation unit that calculates a first value that indicates the degree to which a counselor should listen to or accept the client's utterances based on the estimated level of stress of the client; a second calculation unit that calculates a second value that indicates the degree to which a counselor should empathize with the client's utterances based on the estimated level of self-disclosure of the client; and an instruction generation unit that generates and outputs an instruction to generate a response that indicates attentive listening or acceptance to the client based on the first value and an instruction to generate a response that indicates empathy for the client based on the second value.

[0010] According to the present invention, the degree of listening, acceptance, and empathy for a counselor can be appropriately adjusted according to the state of the counselor.

[0011] Fig. 1 is a diagram illustrating an overview of a response assistance device. Fig. 2 is a diagram illustrating an example of the configuration of a response assistance device. Fig. 3 is a flowchart illustrating an example of a processing procedure executed by the response assistance device. Fig. 4 is a diagram illustrating a specific example of processing executed by the response assistance device. Fig. 5 is a diagram illustrating an example of a computer that executes a response assistance program.

[0012] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a description will be given of an embodiment of the present invention with reference to the drawings, but the present invention is not limited to the embodiment.

[0013] [Overview] An overview of the response support device of this embodiment will be described using Figure 1. First, the response support device acquires information indicating the characteristics of a counselor (e.g., personality traits, diagnosis, etc.) in advance and estimates the characteristics of the counselor ((1)). The response support device also acquires the speech and biosignals of the counselor during counseling and estimates the state of the counselor during counseling ((2)). Then, based on the estimation results obtained in (1) and (2) above, the response support device generates information (e.g., a prompt) instructing the counselor on the level of attentive listening, responsiveness, and empathy that should be shown in a response.

[0014] The information instructing the counselor on the degree of listening, acceptance, and empathy to be shown in a response to the client may be used as reference information when the counselor provides counseling to the client, or may be used as a prompt when the response generation model generates and outputs a response to the client.

[0015] Such a response support device can generate a response in which the degree of attentive listening, acceptance, and empathy is appropriately adjusted to suit the state of the counselee.

[0016] In the following, the response support device will be described taking as an example a case where it generates a prompt for the response generation model described above, and also taking as an example a case where it acquires a biological signal as non-verbal information of a client during counseling.

[0017] [Configuration Example] Next, a configuration example of the response assistance device 10 will be described with reference to Fig. 2. For example, as shown in Fig. 2, the response assistance device 10 includes a storage unit 11, a characteristic parameter conversion unit (third calculation unit) 12, an utterance content acquisition unit 13, a non-verbal information acquisition unit 14, a state estimation unit 15, a first calculation unit 16, a second calculation unit 17, a prompt generation unit (instruction generation unit) 18, a response unit 19, and a response output unit 20.

[0018] [Storage Unit] The storage unit 11 stores data, programs, etc. referenced when the response assistance device 10 executes various processes. For example, the storage unit 11 stores information (characteristic information) indicating personal characteristics of the client, such as personality traits and disease names. The storage unit 11 also stores the client's utterance content acquired by the utterance content acquisition unit 13, non-verbal information acquired by the non-verbal information acquisition unit 14, and responses to the client output by the response output unit 20.

[0019] [Characteristic Parameter Conversion Unit] The characteristic parameter conversion unit 12 acquires characteristic information of the client from the storage unit 11 and converts it into characteristic parameters. The characteristic parameters are parameters that indicate what aspects of the client should be taken into consideration when responding.

[0020] For example, when the characteristic parameter conversion unit 12 acquires a score (AQ score) indicating autistic tendencies of the client as the characteristic information of the client, the characteristic parameter conversion unit 12 converts the AQ score into a characteristic parameter value corresponding to the AQ score by referring to a correspondence table between AQ scores and characteristic parameter values ​​indicated by reference numeral 401 in Fig. 4. The characteristic parameter value here is, for example, a value indicating how much importance should be placed on the emotional and cognitive aspects of the client. Then, the characteristic parameter conversion unit 12 outputs the converted characteristic parameter value (third value).

[0021] In the following description, the value of the characteristic parameter output by the characteristic parameter conversion unit 12 is an AQ score, but is not limited to this.

[0022] For example, consider a case where the client's personal characteristics (disease information) acquired by the characteristic parameter conversion unit 12 include narcissistic personality disorder. In this case, because the client has cognitive distortions, blindly affirming the client may worsen the client's symptoms. Furthermore, denying the client's statements may lead to excessive depression, withdrawal, or even depression. Therefore, when the client's personal characteristics include narcissistic personality disorder, the characteristic parameter conversion unit 12 outputs to the prompt generation unit 18, as a value of the characteristic parameter, for example, "do not deny or affirm the client's incorrect self-perception, but strongly affirm only the events (facts) that the client correctly perceives."

[0023] [Utterance Content Acquisition Unit] Returning to the explanation of Fig. 2, the utterance content acquisition unit 13 acquires the utterance content of the client during counseling. The utterance content may be voice data or text data input by the client.

[0024] [Non-verbal Information Acquisition Unit] The non-verbal information acquisition unit 14 acquires non-verbal information of the client during counseling (for example, EDA (Electrodermal Activity), biosignals such as heart rate, and paralinguistic information extracted from voice data).

[0025] [State Estimation Unit] The state estimation unit 15 estimates the value of a parameter (cognitive parameter) indicating the client's degree of self-disclosure, the value of a parameter (emotion parameter) indicating the client's emotion, and the stress level based on the client's speech content and non-verbal information during counseling.

[0026] For example, the state estimation unit 15 estimates the values ​​of the cognitive parameters and the emotional parameters of the client from the content of the client's speech.

[0027] For example, the state estimation unit 15 estimates the cognitive parameter values ​​and emotional parameter values ​​of the client from the content of the client's utterance using a natural language processing model such as ChatGPT. The state estimation unit 15 also estimates the emotional parameter values ​​and stress level of the client based on non-verbal information (e.g., biological signals) of the client.

[0028] [First Calculation Unit] The first calculation unit 16 calculates a first value (value of the listening / acceptance parameter) indicating the degree to which the counselor should listen to or accept the utterance of the client, based on the level of the client's stress level estimated by the state estimation unit 15. For example, the first calculation unit 16 calculates a higher value of the listening / acceptance parameter as the client's stress level increases. This is because the higher the client's stress level, the more carefully the counselor should listen to and accept the utterance of the client.

[0029] [Second Calculation Unit] The second calculation unit 17 calculates a second value (value of the empathy parameter) indicating the degree to which the counselor should empathize with the speech of the client, based on the value of the client's cognitive parameter estimated by the state estimation unit 15. For example, the second calculation unit 17 sets a higher value of the empathy parameter as the client's degree of self-disclosure decreases. This is because the lower the client's degree of self-disclosure, the more strongly the counselor empathizes with the client, making it easier for the counselor to draw out what the client has to say.

[0030] In addition, the second calculation unit 17 may adjust the value of the empathy parameter depending on whether the emotional state estimated based on the content of the client's utterance matches the emotional state estimated based on the client's non-verbal information.

[0031] For example, as shown by reference numeral 405 in FIG. 4, when the second calculation unit 17 determines that the emotional state (value of the emotional parameter) estimated based on the content of the client's utterance matches the emotional state estimated based on the client's non-verbal information (when the difference in the emotional states is less than a predetermined value), the lower the client's degree of self-disclosure, as described above, the higher the value of the empathy parameter.

[0032] On the other hand, if the second calculation unit 17 determines that the emotional state estimated based on the content of the client's utterance does not match the emotional state estimated based on the client's non-verbal information (if the difference in the emotional states is equal to or greater than a predetermined value), the value of the empathy parameter is set to a predetermined value (e.g., 1) or less, as shown by reference numeral 405 in Figure 4.

[0033] This makes it possible to prevent the second calculation unit 17 from calculating a high value for the empathy parameter for the content of the client's speech when the content of the client's speech is not in line with the client's true feelings, for example.

[0034] [Prompt Generating Unit] Returning to the explanation of Fig. 2, the prompt generating unit 18 generates a prompt sentence (response generating instruction) for generating a response to the content of the utterance of the client.

[0035] For example, the prompt generation unit 18 generates an instruction to generate a response that expresses attentive listening or acceptance of the utterance of the client based on the value of the listening / acceptance parameter output from the first calculation unit 16. The prompt generation unit 18 also generates an instruction to generate a response that expresses empathy for the utterance of the client based on the value of the empathy parameter output from the second calculation unit 17. Furthermore, the prompt generation unit 18 generates an instruction to generate a response that adjusts the balance between emphasis on the emotional and cognitive aspects of the client based on the value of the characteristic parameter output from the characteristic parameter conversion unit 12. The prompt generation unit 18 then outputs these instruction to generate a response to the response unit 19.

[0036] [Response Unit] The response unit 19 generates a response to the client based on the prompt sentence (response generation instruction) generated by the prompt generation unit 18 , and outputs the response to the response output unit 20 .

[0037] For example, the response unit 19 generates and outputs a response to the client by inputting the prompt sentence and the utterance content of the client into a natural language processing model such as ChatGPT. The response output unit 20 outputs the response output from the response unit 19 as text or voice.

[0038] The functions of the characteristic parameter conversion unit 12, the speech content acquisition unit 13, the non-verbal information acquisition unit 14, the state estimation unit 15, the first calculation unit 16, the second calculation unit 17, the prompt generation unit 18, the response unit 19, and the response output unit 20 are realized, for example, by a CPU (Central Processing Unit) executing a program stored in the memory unit 11.

[0039] Furthermore, the response support device 10 may provide the client's doctor or counselor with the content of the client's speech during counseling, non-verbal information, and responses to the client, which are stored in the memory unit 11.

[0040] [Example of Processing Procedure] Next, an example of processing procedure executed by the response support device 10 will be described with reference to Fig. 3. For example, the characteristic parameter conversion unit 12 of the response support device 10 acquires the personal characteristics of the client from the storage unit 11 (S1) and converts them into characteristic parameter values ​​of the client (S2: estimate the value of the characteristic parameter).

[0041] The speech content acquisition unit 13 acquires the speech content of the client during counseling (S3). The non-verbal information acquisition unit 14 acquires the biosignals of the client during counseling (S4). Thereafter, the state estimation unit 15 estimates the cognitive parameter values, emotional parameter values, and stress level of the client based on the speech content of the client acquired in S3 and the non-verbal information of the client acquired in S4 (S5: estimate state parameter values).

[0042] Thereafter, the first calculation unit 16 calculates the value of the listening and acceptance parameter based on the level of the stress level of the client estimated in S5 (S6), and the second calculation unit 17 calculates the value of the empathy parameter based on the level of the cognitive parameter of the client estimated in S5 (S7).

[0043] Thereafter, the prompt generation unit 18 generates a prompt sentence (generation instruction) for generating a response to the content of the client's utterance based on the value of the characteristic parameter estimated in S2, the value of the listening / receptive parameter calculated in S6, and the value of the empathy parameter calculated in S7 (S8: prompt generation).

[0044] Thereafter, the response unit 19 generates and outputs a response to the client by inputting the prompt sentence generated in S8 and the content of the client's utterance into a natural language processing model such as ChatGPT (S9: Response output).

[0045] By executing the above process, the response support device 10 can return a response to the counselor in which the degree of attentive listening, acceptance, and empathy is appropriately adjusted according to the state of the counselor.

[0046] [Specific Example] Next, a specific example of processing executed by the response assistance device 10 will be described with reference to Fig. 4. Here, the response assistance device 10 generates a prompt sentence for generating a response to the content of an utterance by a client, by referring to the information indicated by reference numerals 401 to 406. Note that the information indicated by reference numerals 401 to 406 may be stored in the storage unit 11 of the response assistance device 10, or may be stored in an external device.

[0047] First, the characteristic parameter conversion unit 12 acquires the personal characteristics (characteristic information) of the counseling client from the storage unit 11 (S11) and converts them into characteristic parameters (S12).

[0048] For example, the characteristic parameter conversion unit 12 acquires the AQ (autism tendency score) of the client as the characteristic information of the client, and converts it into a characteristic parameter (for example, 1) by referring to the information indicated by the reference numeral 401 .

[0049] The information indicated by the reference numeral 401 is a correspondence table showing the characteristic parameter values ​​corresponding to the AQ value. In the information indicated by the reference numeral 401, if the AQ value is low ("up to 26"), the characteristic parameter value is 2 (emphasis is placed on the emotional aspect of the client), if the AQ value is medium ("27 to 32"), the characteristic parameter value is 1 (emphasis is placed on the cognitive aspect of the client while also taking into account the emotional aspect), and if the AQ value is high ("33 or above"), the characteristic parameter value is 3 (emphasis is placed on the cognitive aspect of the client).

[0050] The characteristic parameter conversion unit 12 refers to the information indicated by the symbol 401, and if the client has a low tendency toward autism, converts the information into a characteristic parameter value that emphasizes the client's emotional aspect, and if the client has a high tendency toward autism, converts the information into a characteristic parameter value that emphasizes the client's cognitive aspect (rather than the emotional aspect).

[0051] Furthermore, when the utterance content acquisition unit 13 acquires the utterance content of the client (S21), the state estimation unit 15 estimates the values ​​of the cognitive parameters and the emotional parameters from the utterance content acquired in S21 (S22).

[0052] The state estimation unit 15 estimates the value of the cognitive parameter (degree of self-disclosure) of the client from, for example, the amount of explanation about the client himself / herself in the content of the client's utterance.

[0053] For example, the state estimation unit 15 quantifies the amount of explanation (number of words) about the past and present states of the client that begin with "I" (or have no subject) in the utterance content of the client. Then, the state estimation unit 15 estimates the degree of self-disclosure of the client by referring to the quantified amount of explanation about the past and present states and the information indicated by reference numeral 402.

[0054] The information indicated by the reference numeral 402 is information indicating the degree of self-disclosure corresponding to the number of words per unit time (for example, 2 minutes).

[0055] In this information, the greater the number of words per unit time, the higher the self-disclosure level is set. For example, in the information indicated by reference numeral 402, if the number of words per unit time is low ("0 to 30"), the self-disclosure level is set to 1, if the number of words per unit time is medium ("31 to 60"), the self-disclosure level is set to 2, and if the number of words per unit time is high ("61 or above"), the self-disclosure level is set to 3.

[0056] Furthermore, the state estimation unit 15 estimates the value of the emotion parameter from the utterance content acquired in S21, for example.

[0057] For example, consider a case where the utterance content acquisition unit 13 acquires an utterance from the client saying, "I felt very frustrated at that time." In this case, the state estimation unit 15 extracts the emotion-expressing word "frustrating" (emotion word) from the utterance content. Then, the state estimation unit 15 refers to a DB (database) that indicates the emotion category for each emotion word, and determines the emotion category (e.g., negative) of the emotion word "frustrating."

[0058] Next, the state estimation unit 15 references the information indicated by the reference numeral 403 and estimates the values ​​of the emotion parameters of the above emotion categories.

[0059] The information indicated by the reference numeral 403 indicates the value of an emotion parameter corresponding to an emotion category. For example, the state estimation unit 15 refers to the information indicated by the reference numeral 403 and estimates the emotion parameter value of "-1" for the emotion category (negative) of "frustrated."

[0060] The non-verbal information acquisition unit 14 also acquires the client's biometric signals (e.g., heart rate) (S31). The state estimation unit 15 then extracts biometric features (e.g., heart rate fluctuations) from the biometric signals within a predetermined time period (e.g., 5 minutes) (S32). Thereafter, the state estimation unit 15 estimates the client's emotional parameter values ​​and stress level from the biometric features extracted in S32 (S33).

[0061] For example, if the state estimation unit 15 estimates that the emotion category of the client is "negative" based on the biometric feature of the client, the state estimation unit 15 references the information indicated by the reference numeral 403 and estimates the value of the emotion parameter corresponding to the emotion category (e.g., "-1"). The state estimation unit 15 also estimates the stress level (e.g., "3") from the biometric feature of the client. For the above estimation, for example, information (not shown) that associates the biometric feature with the emotion category value and the stress level is referenced.

[0062] After S33, the first calculation unit 16 calculates the value of the listening and acceptance parameter for the client based on the stress level of the client estimated in S33 (S41).

[0063] For example, the first calculation unit 16 calculates the value of the listening and acceptance parameter based on the information indicated by the reference numeral 404. The information indicated by the reference numeral 404 indicates the value of the listening and acceptance parameter corresponding to the stress level. In this information, the higher the stress level of the client, the higher the value of the listening and acceptance parameter is set.

[0064] For example, the information indicated by the symbol 404 indicates that when the stress level is the lowest value "1", the value of the listening / accepting parameter is set to "0 (responds at the default frequency)", when the stress level is the medium value "2", the value of the listening / accepting parameter is set to "1 (responds 1.5 times as frequently as the default)", and when the stress level is the highest value "3", the value of the listening / accepting parameter is set to "2 (responds twice as frequently as the default)".

[0065] Furthermore, the second calculation unit 17 calculates the value of the empathy parameter after a predetermined time (e.g., 5 minutes) has elapsed since the client started speaking (S42). For example, the second calculation unit 17 calculates the value of the empathy parameter based on the value of the cognitive parameter (degree of self-disclosure) and the value of the emotion parameter of the client estimated from the content of the utterance in S22, and the value of the emotion parameter of the client estimated from the biometric feature in S33.

[0066] For example, the second calculation unit 17 calculates the value of the empathy parameter by referring to the information indicated by the reference numeral 405 .

[0067] The information indicated by the reference numeral 405 is information indicating the value of the empathy parameter corresponding to the value of the client's cognitive parameter (degree of self-disclosure). For example, the lower the degree of self-disclosure, the higher the empathy parameter value is set. Furthermore, for example, if the client's emotional parameter value estimated from the content of the utterance does not match the emotional parameter value estimated from the biometric features, the empathy parameter value is set to a predetermined value or less.

[0068] For example, consider a case where the value of the client's cognitive parameter (degree of self-disclosure) is "3," the client's emotional parameter estimated from the content of the utterance is "-1," and the emotional parameter of the client estimated from the biometric feature is "-1." In this case, the emotional parameter value estimated from the content of the utterance matches the emotional parameter value estimated from the biometric feature. The degree of self-disclosure is also "3." Therefore, the second calculation unit 17 outputs "0," which corresponds to the degree of self-disclosure of "3" in the "Emotional parameter = Match" column in the information indicated by the reference numeral 404, as the value of the empathy parameter.

[0069] If the value of the emotion parameter estimated from the content of the utterance does not match the value of the emotion parameter estimated from the biometric feature, the second calculation unit 17 outputs the value corresponding to the degree of self-disclosure of the client in the column “Emotion parameter=Mismatch” in the information indicated by the reference numeral 404 as the value of the empathy parameter.

[0070] Thereafter, the prompt generating unit 18 generates a prompt instructing the client to respond based on the value of the listening / receptive parameter calculated in S41 and the value of the empathy parameter calculated in S42 (S43).

[0071] For example, if the value of the listening / acceptance parameter calculated in S41 is "2 (responding twice as frequently as the default)," then the accepting / listening response to the client will be made twice as frequently as the default. For example, if the default frequency of accepting / listening is once every 10 sentences spoken by the client, then twice the frequency is once every five sentences spoken by the client. Therefore, the prompt generation unit 18 generates the following prompt, which instructs the client to make an accepting / listening response once every five sentences spoken by the client:

[0072] Up to the fifth sentence: "Please do not respond by nodding, listening, or offering specific responses that show acceptance." From the fifth sentence onwards: "Please respond by listening to what the other person is saying and showing acceptance."

[0073] The prompt generator 18 also generates an empathy prompt by referring to information indicating the content of the prompt corresponding to the empathy parameter value (see reference numeral 406). For example, as shown by reference numeral 406, the information is set so that the content of the prompt indicates a higher level of empathy as the empathy parameter value increases.

[0074] For example, in the information indicated by the reference numeral 406, if the empathy parameter value is "0", "no empathy prompt" is set, and if the empathy parameter value is "1", a prompt instructing "a brief summary of the other party's (consultant's) condition" is set. Also, if the empathy parameter value is "2", a prompt instructing "a detailed summary of the other party's (consultant's) condition" is set, and if the empathy parameter value is "3", a prompt instructing "a brief summary of the other party's (consultant's) condition and an empathic response" is set.

[0075] For example, if the value of the empathy parameter calculated in S42 is "0," the prompt generation unit 18 does not generate an empathy prompt by referring to the information indicated by the reference numeral 406. If the characteristic parameter converted in S12 is "1," the prompt generation unit 18 refers to the information indicated by the reference numeral 406 and generates a prompt sentence that instructs a response that emphasizes the cognitive aspect of the client while also taking into account their emotional aspect.

[0076] The response unit 19 then inputs the prompt generated by the prompt generation unit 18 and the utterance content of the client into the response generation model, and generates a response to the utterance content of the client. The response output unit 20 then outputs the response generated by the response unit 19.

[0077] By executing the above processing, the response support device 10 can return to the counselor a response in which the degree of attentive listening, acceptance, and empathy is appropriately adjusted according to the counselor's state. As a result, more effective counseling and intervention for the counselor than existing technologies is possible, and a response can be made that elicits the client's main complaint and improves it. Furthermore, the response support device 10 can store the client's utterances, biological reactions, response results, etc. during counseling and provide them to the client's doctor or professional counselor, thereby supporting the counseling provided by the doctor or professional counselor.

[0078] [System Configuration, etc.] The components of each unit shown in the figure are conceptual functional units and do not necessarily have to be physically configured as shown. In other words, the specific form of distribution and integration of each device is not limited to that shown, and all or part of them can be functionally or physically distributed and integrated in any unit depending on various loads, usage conditions, etc. Furthermore, all or any part of the processing functions performed by each device can be realized by a CPU and a program executed by the CPU, or can be realized as hardware using wired logic.

[0079] Furthermore, among the processes described in the above embodiments, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically using a known method.In addition, the information including the processing procedures, control procedures, specific names, various data and parameters shown in the above documents and drawings can be changed as desired unless otherwise specified.

[0080] [Program] The response assistance device 10 can be implemented by installing a program (response assistance program) as package software or online software on a desired computer. For example, by executing the program on the response assistance device, the response assistance device can function as the response assistance device 10. The response assistance device referred to here includes mobile communication terminals such as smartphones, mobile phones, and PHS (Personal Handyphone Systems), as well as terminals such as PDAs (Personal Digital Assistants).

[0081] 5 is a diagram showing an example of a computer that executes a response assistance program. Computer 1000 includes, for example, memory 1010 and CPU 1020. Computer 1000 also includes a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface 1070. These components are connected by a bus 1080.

[0082] The memory 1010 includes a read-only memory (ROM) 1011 and a random access memory (RAM) 1012. The ROM 1011 stores a boot program such as a basic input / output system (BIOS). The hard disk drive interface 1030 is connected to a hard disk drive 1090. The disk drive interface 1040 is connected to a disk drive 1100. A removable storage medium such as a magnetic disk or optical disk is inserted into the disk drive 1100. The serial port interface 1050 is connected to a mouse 1110 and a keyboard 1120, for example. The video adapter 1060 is connected to a display 1130, for example.

[0083] The hard disk drive 1090 stores, for example, an OS 1091, an application program 1092, a program module 1093, and program data 1094. That is, the programs that define the processes executed by the response assistance device 10 are implemented as program modules 1093 in which computer-executable code is written. The program modules 1093 are stored, for example, in the hard disk drive 1090. For example, the program modules 1093 for executing processes similar to those of the functional configuration of the response assistance device 10 are stored in the hard disk drive 1090. The hard disk drive 1090 may be replaced with an SSD (Solid State Drive).

[0084] Data used in the processing of the above-described embodiment is stored as program data 1094, for example, in the memory 1010 or the hard disk drive 1090. The CPU 1020 then reads the program module 1093 or the program data 1094 stored in the memory 1010 or the hard disk drive 1090 into the RAM 1012 as necessary and executes them.

[0085] The program module 1093 and program data 1094 may not necessarily be stored in the hard disk drive 1090, but may also be stored in a removable storage medium and read by the CPU 1020 via the disk drive 1100 or the like. Alternatively, the program module 1093 and program data 1094 may be stored in another computer connected via a network (such as a local area network (LAN) or a wide area network (WAN)). The program module 1093 and program data 1094 may then be read by the CPU 1020 from the other computer via the network interface 1070.

[0086] REFERENCE SIGNS LIST 10 Response support device 11 Storage unit 12 Characteristic parameter conversion unit 13 Speech content acquisition unit 14 Non-verbal information acquisition unit 15 State estimation unit 16 First calculation unit 17 Second calculation unit 18 Prompt generation unit 19 Response unit 20 Response output unit

Claims

1. A response support device comprising: a state estimation unit that estimates a level of self-disclosure and stress level of a client based on the content of the client's speech and non-verbal information during counseling; a first calculation unit that calculates a first value indicating the degree to which a counselor should listen to or accept the client's speech based on the estimated level of stress level of the client; a second calculation unit that calculates a second value indicating the degree to which a counselor should empathize with the client's speech based on the estimated level of self-disclosure of the client; and an instruction generation unit that generates and outputs an instruction to generate a response indicating attentive listening or acceptance to the client based on the first value, and an instruction to generate a response indicating empathy for the client based on the second value.

2. The response support device according to claim 1, further comprising a third calculation unit that calculates a third value indicating what aspects of the client should be taken into consideration when responding based on personal characteristics indicating the client's personality or illness, and the instruction generation unit further generates and outputs instructions for generating a response to the client based on the third value.

3. The response support device described in claim 2, characterized in that the third calculation unit calculates the third value indicating the degree of emphasis to be placed on the emotional and cognitive aspects of the client based on the personal characteristics that indicate the client's autistic tendencies, and the instruction generation unit generates and outputs instructions for generating a response to the client that adjusts the balance of emphasis on the emotional and cognitive aspects of the client based on the third value.

4. The response support device of claim 1, characterized in that the first calculation unit calculates the first value higher the higher the stress level of the client, and the second calculation unit calculates the second value higher the lower the client's level of self-disclosure.

5. The response support device according to claim 4, characterized in that the state estimation unit further estimates the emotional state based on the content of the client's utterance and the emotional state based on non-verbal information of the client, and the second calculation unit sets the second value to a predetermined value or less when the emotional state estimated based on the content of the client's utterance does not match the emotional state estimated based on the non-verbal information of the client.

6. The response support device according to claim 5, further comprising: a response generation unit that generates a response to the client by inputting an instruction to generate the response to the client into a model that generates the response to the client; and a response output unit that outputs the generated response to the client.

7. A response support method executed by a response support device, comprising the steps of: estimating a level of self-disclosure and stress level of a client based on the content of the client's utterances and non-verbal information during counseling; calculating a first value indicating the degree to which a counselor should listen to or accept the client's utterances based on the estimated level of stress level of the client; calculating a second value indicating the degree to which a counselor should empathize with the client's utterances based on the estimated level of self-disclosure of the client; and generating and outputting an instruction to generate a response indicating attentive listening or acceptance to the client based on the first value, and an instruction to generate a response indicating empathy for the client based on the second value.

8. A response support program for causing a computer to execute the following steps: a step of estimating the level of self-disclosure and stress of a client based on the content of the client's speech and non-verbal information during counseling; a step of calculating a first value indicating the degree to which the counselor should listen to or accept the client's speech based on the estimated level of stress of the client; a step of calculating a second value indicating the degree to which the counselor should empathize with the client's speech based on the estimated level of self-disclosure of the client; and a step of generating and outputting an instruction to generate a response indicating attentive listening or acceptance to the client based on the first value, and an instruction to generate a response indicating empathy for the client based on the second value.

Images