Communication robot

Abstract

To reduce fraud damage more effectively when a robot that snuggles up to the elderly is used than when the robot is not used.SOLUTION: The communication robot includes dialogue means for dialoguing with a user, call information acquisition means for acquiring information on a call made by the user with a third party, risk degree determination means for determining the degree of risk in which the call is fraudulent by using the acquired call information, characteristic acquisition means for acquiring information on individual characteristics of the user, wherein the dialogue means includes output means for determining and outputting the content of dialogue based on the information on the individual characteristics when the degree of risk is determined to be high.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to a communication robot. [Background technology]

[0002] Patent document 1 describes a server device that includes a memory processing unit that stores conversation information regarding a conversation between a first telephone and a second telephone in a memory unit, a counting unit that retrieves the conversation information from the memory unit every time a predetermined time has elapsed since voice communication was established and counts the number of specific words contained in the conversation, a judgment unit that judges the level of danger for the conversation based on the counted number, and a transmission unit that, if the level of danger exceeds a first predetermined value, sends an email indicating that a special fraud is occurring to a mobile terminal related to the user, and, if the level of danger exceeds a second predetermined value that is greater than the first predetermined value, sends instruction information to an exchange related to the voice communication to disconnect the voice communication between the first telephone and the second telephone. Patent document 2 describes a suspicious call discrimination program that includes a call content acquisition step for acquiring the call content, an affiliation information extraction step for extracting the affiliation information claimed by the other party from the call content, an affiliation determination step for determining whether the affiliation information contains incorrect information by referring to an organization database in which organization information related to organizations is registered, a call content determination step for determining whether the call content contains NG keywords by referring to a NG keyword database in which words used in calls from impostors are registered if the affiliation information does not contain incorrect information, and a notification step for notifying that the call is suspicious if the affiliation information contains incorrect information or if the call content contains NG keywords. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-036231 A [Patent Document 2] JP 2007-228384 A Summary of the Invention [Problem to be solved by the invention]

[0004] Thus, in the prior art, a technique for issuing a warning to the elderly has been proposed to deal with special frauds. However, even if a warning is issued, the elderly who receive a special fraud call may become upset, for example, and may be preoccupied with the scammer's story, making it impossible for them to accept the warning. In such cases, it is desirable for the elderly person's family or acquaintances to persuade them to change their way of thinking. However, since fraudsters generally try to induce the elderly person not to contact others, in reality, persuasion by family or acquaintances is not carried out, and fraud victimization cannot be prevented. In recent years, there are communication robots that are placed at the user's hand and communicate with the user. These communication robots can be as close to the user as a pet, and can become closer to the user than a family member. By using a robot that is close to the elderly, one of the users, it is expected that fraud victims of the elderly can be more effectively prevented. The present invention aims to reduce fraud damage more effectively than if a robot that supports the elderly were not used. [Means for solving the problem]

[0005] The invention described in claim 1 comprises: a dialogue means for dialogue with a user; a call information acquisition means for acquiring call information between the user and a third party; a risk determination means for determining a degree of risk that the call is fraudulent using the acquired call information; and a characteristic acquisition means for acquiring information regarding individual characteristics of the user. and an emotion information acquisition means for acquiring information regarding the emotion of the user during or after a call, The dialogue means, when it is determined that the degree of risk is high, transmits information regarding the individual characteristics to the and information about the emotion. Based on this, the dialogue content is determined and output the information on emotion is information indicating a degree of upset of the user, determined based on at least one of voice information of the user and a facial expression of the user, which are included in the call information; and the information indicating the degree of upset is information obtained by comparing history information on a dialogue between the user and the dialogue means and information on the emotion acquired by the emotion information acquisition means. It is a communication robot. The invention described in claim 2 further comprises the information on each of the individual characteristics, The aboveAs understood from historical information, The above The communication robot according to claim 1, characterized in that it is a characteristic behavioral pattern possessed by a user. The invention described in claim 3 is the communication robot described in claim 2, characterized in that the dialogue means changes at least one of the contents of the voice and the characteristics of the voice from the stored voice information in accordance with the individual characteristic behavioral patterns of the user, or uses new voice information to dialogue with the user. 。 request request 4 The invention described in is characterized in that the information about the individual characteristics is information about an event that has generated lively dialogue with the user, Dialogue The means 、 The communication robot according to claim 1, characterized in that the content of the conversation is determined based on information regarding matters that have led to lively conversation with the user. Effect of the Invention

[0006] According to the present invention, it is possible to more effectively reduce fraud damage compared to not using a robot that supports the elderly. [Brief description of the drawings]

[0007] [Figure 1] 1 is a diagram showing a hardware configuration of a robot warning system to which an embodiment of the present invention is applied. [Diagram 2] FIG. 2 is a diagram illustrating a hardware configuration of a communication robot. [Diagram 3] FIG. 2 is a diagram illustrating a hardware configuration of a server. [Figure 4] FIG. 2 is a diagram for explaining functions of a communication robot. [Diagram 5] FIG. 2 is a diagram for explaining the functions of a server. [Figure 6-1] 13 is a flowchart showing a process flow when the communication robot issues a warning. [Figure 6-2]11 is a flowchart showing a process for determining a risk of fraud. [Figure 6-3] 13 is a flowchart showing a process for recognizing a user's emotion. [Figure 7] FIG. 13 is a diagram showing an example of a table used when determining the content of a dialogue and the manner of voice. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] [Overall configuration of the robot warning system] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. FIG. 1 is a diagram showing a hardware configuration of a robot warning system 1 to which the present embodiment is applied. The robot warning system 1 according to the present embodiment includes a communication robot 10 that communicates with an elderly person, who is an example of a user. When the elderly person answers a phone call that is suspected to be a fraud such as a specialized fraud, the communication robot 10 issues a warning to the elderly person. Here, the "specialized fraud" refers to a crime in which a perpetrator makes a phone call to the elderly person to gain their trust without meeting them face to face, and then defrauds them of cash or the like by transferring the money to a specified savings account or by other means. There are various methods of specialized fraud, and new methods are constantly being developed.

[0009] The robot warning system 1 according to the present embodiment includes a communication robot 10, a telephone 70 in the elderly person's home, and a server 30 that provides information to the communication robot 10. The communication robot 10 and the server 30 are connected via a network 50. The robot warning system 1 also includes a connection device 60 for connecting the communication robot 10 to the network 50. The connection device 60 is not particularly limited, but may be, for example, a Wi-Fi connection device. However, if it is difficult to prepare a Wi-Fi connection environment in the elderly person's home, a device using, for example, a Subscriber Identity Module Card (SIM card) for data communication may be used as the connection device 60. In such a case, the connection to the network 50 is made via a predetermined base station. The connection device 60 may also be connected to the telephone 70, and the communication robot 10 may be configured to be connected to the telephone 70 via the connection device 60.

[0010] Furthermore, the robot warning system 1 includes a terminal 80 owned by a person related to the user, such as a mobile phone or smartphone owned by the child of the elderly person, and a database 40 in which information on special fraud is accumulated. The terminal 80 owned by a person related to the user may be owned by a relative such as the child or grandchild of the elderly person, or may be a terminal owned by a caregiver, for example. The terminal 80 owned by the person related to the user and the server 30 are connected via a network 50.

[0011] [Hardware configuration of communication robot 10] 2 is a diagram showing a hardware configuration of the communication robot 10. The communication robot 10 includes a control unit 11 configured with a CPU (Central Processing Unit) for controlling the entire communication robot, a memory 12 such as a RAM (Random Access Memory) used as a working area for calculations, and a storage unit 13 which is a storage device such as a semiconductor memory used for storing programs and various setting data. The communication robot 10 also includes a communication unit 14 which is connected to a network 50 via a connection device 60 and transmits and receives information to and from a server 30, an operation unit 15 such as a button or a touch panel which receives operations from a user to the communication robot 10, a display unit 16 which is configured with a liquid crystal display or the like which displays images, text information, and the like to the user, and a display control unit 17 which controls the display unit 16. The communication robot 10 also includes an audio output unit 18 which is configured with a speaker or the like which outputs audio to the user, an audio acquisition unit 19 which is configured with a microphone or the like which records the audio uttered by the user, and an image acquisition unit 20 which is configured with a camera or the like which captures the user's appearance and facial expression and acquires them as still or moving image information.

[0012] [Hardware configuration of server 30] FIG. 3 is a diagram showing a hardware configuration of the server 30. As shown in FIG. The server 30 is realized by a computer. The server 30 has a control unit 31 composed of a CPU (Central Processing Unit) that controls the entire device, a memory 32 such as a RAM (Random Access Memory) that is used as a work area during calculations, and a storage unit 33 that is a storage device such as a HDD (Hard Disk Drive) or a semiconductor memory that is used to store programs and various setting data. The server 30 also has a communication unit 34 that transmits and receives data via a network 50. The server 30 also has an operation unit 35 such as a keyboard, a pointing device, a touch panel, etc. that accepts input operations from a server administrator that manages the server on the server 30 side, a display unit 36 ​​composed of a liquid crystal display or the like that displays images, text information, etc. to the server administrator, and a display control unit 37 that controls the display unit 36.

[0013] [Functional configuration of communication robot 10] Next, the function of the communication robot 10 will be described with reference to FIG. FIG. 4 is a diagram for explaining the functions of the communication robot 10. As shown in FIG. The control unit 11 of the communication robot 10 processes information input from the communication unit 14 , the voice acquisition unit 19 , and the image acquisition unit 20 , and outputs the processed information from the communication unit 14 and the voice output unit 18 . Furthermore, the processing performed by the control unit 11 acquires information stored in the storage unit 13 according to the processing content, and stores the information in the storage unit 13 according to the processing content. The information stored in the storage unit 13 includes "call information", "telephone directory information", "dangerous word information", "crime prevention information", "dialogue history information", "user characteristic information", and "dialogue scenario information".

[0014] Here, "call information" refers to information about conversations the user has with the communication robot 10, or conversations the user has with other people. Moreover, "telephone directory information" refers to information on telephone numbers registered in telephone set 70, information on the frequency with which the user makes calls, and the like. Furthermore, "dangerous word information" refers to information on "phrases" and "keywords" used in specialized fraud. Examples of information on "phrases" and "keywords" used in specialized fraud include "your phone number has changed" and "you need to check for misused cards." The "crime prevention information" is crime prevention information provided to the user when the communication robot 10 has a daily conversation with the user. For example, the "crime prevention information" is information that the communication robot 10 tells the user during daily conversation, such as "there is a scam going around where people pretend to be police officers to get cash card numbers," and is information about popular scam methods and countermeasures against scams. Moreover, the "dialogue history information" is information regarding a dialogue between a user and the communication robot 10. For example, it is voice information of the user in a dialogue between the user and the communication robot 10. The information of the "dialogue history" is an example of history information regarding a dialogue. The "user characteristic information" is information about characteristic actions and words uttered by the user during a conversation with the communication robot 10. For example, the "user's personality" and "what the user likes" are examples of user characteristic information. It is the content of a conversation that the user finds enjoyable. Moreover, the "dialogue scenario information" is information regarding a scenario when the communication robot 10 dialogues with a user. For example, multiple scenarios are prepared according to the characteristics of the user for the "dialogue scenario information". The "dialogue scenario information" sets the words and actions that the communication robot 10 will utter in response to the words and actions uttered by the user.

[0015] The control unit 11 mainly includes a call information processing unit 111 that processes call information acquired by the voice acquisition unit 19, and a communication information input unit 112 that stores information input from the communication unit 14 in the storage unit 13. The control unit 11 also includes a risk determination unit 113 that determines the risk that the call answered by the user is a fraudulent call, an emotion recognition unit 114 that recognizes the user's emotion, and a user characteristic recognition unit 115 that recognizes the user's characteristics. The control unit 11 also includes a persuasion information generation unit 116 that generates information for persuading the user when there is a risk of fraud against the user, and a communication information output unit 117 that outputs information to the server 30 using the communication unit 14.

[0016] The call information processing unit 111 processes the call information among the input information. The process performed on the call information is, for example, a voice recognition process for the voice during the call. An example of the voice recognition process is a process of converting the voice information into text data. The call information processing unit 111 functions as an example of a call information acquisition unit.

[0017] The communication information input unit 112 processes the information acquired from the communication unit 14 and inputs the information to each functional unit. The information acquired from the communication unit 14 includes information acquired from the server 30 and information acquired from the telephone set 70. Examples of information acquired from the server 30 include "dangerous word information," "dialogue scenario information," and "crime prevention information." Examples of information obtained from telephone 70 include information regarding the telephone number of the other party and the time of the call, voice information of the user recorded by a microphone installed in telephone 70, and voice information of the other party obtained by telephone 70.

[0018] The risk determination unit 113 determines the level of risk that the call answered by the user is a fraudulent call. The risk determination unit 113 determines the level of risk using, for example, "call information," "phonebook information," and "dangerous word information" stored in the storage unit 13. For example, if the phone number of a family member or friend with whom the user frequently calls matches the phone number of the callee, the risk is determined to be low. Also, for example, if there is voice information that is the same as or similar to the dangerous phrase "the phone number has changed" in a call made by the user with a third party, the risk is determined to be high.

[0019] The emotion recognition unit 114 recognizes the emotion of the user using the voice information of the user included in the call information and the image information of the user's facial expression acquired from the image acquisition unit 20. The degree of the user's agitation is an example of the emotion of the user. For example, the emotion recognition unit 114 extracts the characteristics of the user's voice during a call and performs emotion recognition by comparing the characteristics of the user's voice included in the user's characteristics extracted by the user characteristic recognition unit 115 described later. For example, if the "words often used when the user is agitated" are "Why?", when the word "Why?" is used during a call, the emotion recognition unit 114 recognizes that the user is agitated. Furthermore, the degree of agitation can be determined based on the number of times the "words often used when the user is agitated" appear during a call, the magnitude of the difference between the volume of the user's voice when speaking normally and the volume of the voice during a call, and the like. Another example is that when a user does not behave in a way that he or she would normally do when feeling at ease, the emotion recognition unit 114 recognizes that the user is "feeling anxious." Examples of the features of the user's voice include the user's speaking speed, the volume of the voice, etc. Furthermore, the user's emotion may be recognized using image information of the user's facial expression captured by the image acquisition unit 20. Emotion recognition unit 114 is an example of an emotion information acquisition means.

[0020] The user characteristic recognition unit 115 recognizes the characteristics of the user from the "dialogue history information" stored in the memory unit 13, and stores the recognized characteristics in the memory unit 13 as user characteristics. In addition, the user characteristic recognition unit 115 extracts, for example, words that the user frequently uses in normal times from the text information of the "dialogue history information", and stores the words that the user frequently uses in normal times in the memory unit 13 as user characteristics. Actions and words uttered by the user while interacting with the communication robot 10 are an example of a characteristic behavioral pattern possessed by the user. Information on the user characteristics is an example of information on individual characteristics. In addition, the user characteristic recognition unit 115 stores the contents of lively conversations in dialogue with the user in the storage unit 13. The contents of lively conversations in dialogue with the user are extracted and stored in the storage unit 13 as the user's favorite objects. The user characteristic recognition unit 115 functions as an example of a characteristic acquisition means.

[0021] The persuasive information generating unit 116 determines the content and manner of the voice to be output to the user when there is a risk of fraud in a call with a third party, and generates persuasive information for persuading the user. The content and manner of the voice are determined based on the characteristics of the user, and persuasive information according to the characteristics of the user is generated. For example, when the user has a personality that is "emotionally unstable," persuasive information is generated using a dialogue scenario created for a person with a personality that is "emotionally unstable." A robot that is close to the user extracts the user's characteristics from daily conversations and generates persuasive information according to the user's characteristics, thereby generating persuasive information that is easily accepted by the user. The persuasive information generating unit 116 may generate persuasive information based on the user's emotions. For example, when the user is in a hurry, the persuasive information generating unit 116 generates persuasive information in a voice format that calms the user by slowing down the playback speed of the voice to be played back. The persuasive information generating unit 116 functions as an example of a dialogue means and an output means.

[0022] The communication information output unit 117 outputs information to the server 30 using the communication unit 14. In order to inform a person related to the user that a call that is likely to be fraudulent has been received, the communication information output unit 117 requests the server 30 to notify the terminal 80 of the person related to the user. Furthermore, the communication information output unit 117 periodically outputs information to the server 30 requesting connection to the server 30. The communication robot 10 periodically connects to the server 30 and acquires information from the server 30. An example of a periodic connection is a connection at a predetermined interval, such as once a day. By acquiring information from the server 30 via the network 50 and updating the information stored in the storage unit 13, information on new fraud methods can be obtained. Note that instead of periodically connecting to the server 30, a configuration may be adopted in which the server 30 is constantly connected, and a configuration may be adopted in which when information is updated in the server 30, information is transmitted to the communication robot 10 at the discretion of the server 30.

[0023] [Functional configuration of server 30] Next, the function of the server 30 will be described with reference to FIG. FIG. 5 is a diagram for explaining the functions of the server 30. As shown in FIG. The control unit 31 processes information input from the communication unit 34 and the operation unit 35, and outputs the processed information from the communication unit 34. Furthermore, the control unit 31 acquires information stored in the storage unit 33 according to the processing contents, and also stores information generated according to the processing contents in the storage unit 33. The information stored in the storage unit 33 includes "dangerous word information", "dialogue scenario information", "user-related information", and "crime prevention information". Here, "user-related information" refers to information about the user, such as the user's age and gender. This user-related information may also include contact information for related parties of the user. The contact information for related parties of the user is, for example, an emergency contact to be contacted if the user receives a high-risk phone call from a fraudulent person, such as a son, grandchild, or a person related to the facility.

[0024] The control unit 31 includes a communication information input unit 311 that stores information received from the communication unit 34 in the memory unit 33, and a communication information output unit 315 that outputs information from the communication unit 34. It also includes a fraud information processing unit 312 that processes acquired information related to fraud, a fraud risk processing unit 313 that performs various processes related to fraud, such as setting a risk level for phrases and words that are often used in specialized frauds, and a user-related information processing unit 314 that performs various processes related to information related to the user, such as updating or changing contact information.

[0025] The communication information input unit 311 distributes information input from the communication unit 34 to each functional unit and stores the information in the memory unit 33. The information input from the communication unit 34 includes information on specialized fraud obtained from a database 40. The database 40 is, for example, a database that accumulates data on specialized fraud created by an administrative agency such as the police, or a database created by a service provider of the robot warning system 1. The information input from the communication unit 34 includes a request for information transmission from the communication robot 10, a request for notification to a terminal 80 held by a related person of the user, and the like.

[0026] The fraud information processing unit 312 analyzes information relating to fraud obtained from the database 40 via the communication unit 34, extracts phrases and keywords that are often used in fraud, and stores them as dangerous word information in the memory unit 33. In addition, if there is new information relating to fraud methods and fraud countermeasures, it stores it in the memory unit 33 as updated crime prevention information.

[0027] The fraud risk processing unit 313 associates a risk level with the phrases and keywords extracted by the fraud information processing unit 312. For example, if a certain phrase is found during a call and there is a high possibility that the call is related to a special fraud, the risk level for that phrase is set to a high value. The level of risk may be set based on machine learning using AI, for example, or may be set by a server administrator.

[0028] The communication information output unit 315 outputs information to the communication robot 10 and the terminal 80 of a person related to the user via the communication unit 34. Information output to the communication robot 10 includes "dangerous word information," "crime prevention information," "dialogue scenario information," and the like. The communication information output unit 315 may select and transmit information suitable for the user using user-related information. For example, if the user is elderly, it may select a "dialogue scenario" created for elderly people from all "dialogue scenarios" and transmit it to the communication robot 10. The information output to the terminal 80 of the user's related party is information notifying the user that he or she has received a phone call that is likely to be fraudulent.

[0029] When information is input from the server administrator via the operation unit 35, the control unit 31 edits the information stored in the storage unit 33 in response to the operation of the server administrator. For example, when the server administrator learns of a new fraud method, the server administrator adds a phrase used in the new fraud method to the list of danger words. This allows the server administrator to quickly prepare for new frauds.

[0030] [Processing performed by communication robot 10] Next, the processing performed by the communication robot 10 will be described with reference to FIGS. 4, 6-1, 6-2, 6-3, and 7. FIG. FIG. 6-1 is a flowchart showing a process flow when the communication robot 10 issues a warning. First, the risk determination unit 113 performs a risk determination process for determining the risk that a call between a user and a third party is related to a specialized fraud (step 1001). The risk determination process will be described in detail with reference to FIG. 6-2. If the risk determination process determines that there is a risk (YES in step 1002), the communication information output unit 117 outputs a request to the server 30 to notify people related to the user, such as the user's children or grandchildren, that the user is in danger (step 1003). When the communication information input unit 311 (see FIG. 5) of the server 30 accepts the request in step 1003, the communication information output unit 315 (see FIG. 5) notifies the terminal of the user's elderly son or daughter that the user has received a phone call that may be a special fraud.

[0031] Thereafter, the emotion recognition unit 114 performs emotion recognition processing to recognize the emotion of the user (step 1004). The emotion recognition processing will be described in detail with reference to FIG. 6-3. The persuasive information generation unit 116 determines the dialogue content and the voice manner based on the user's characteristics and the user's emotion recognized by the emotion recognition unit 114 (step 1005). The persuasive information generation unit 116 sends playback information based on the dialogue content and voice manner determined in step 1005 to the voice output unit 18. The voice output unit 18 converses with the user about the determined dialogue content in the set voice manner (step 1006).

[0032] Next, the risk determination process in step 1001 will be described with reference to FIG. FIG. 6-2 is a flowchart showing the process of determining the risk of fraud. When a dangerous word is found during a call, the danger level determination unit 113 shown in FIG. 4 first sets 0 to the danger level X as a variable for accumulating the danger level (step 2001). Next, it is determined whether or not the voice-recognized text contains a dangerous word (step 2002). More specifically, the danger level determination unit 113 compares the text extracted by the call information processing unit 111 through voice recognition with the dangerous word information stored in the storage unit 13 to determine whether or not the dangerous word is present. Here, the information that the call information processing unit 111 converts into text can include not only the voice of the user during the call but also the voice information of the other party. When it is determined that the voice-recognized text contains a dangerous word (YES in step 2002), the danger level corresponding to the dangerous word found in the text is added to the danger level X (step 2003). The dangerous word information stored in the storage unit 13 is used as the danger level corresponding to the dangerous word. When it is not determined in step 2002 that the voice-recognized text contains a dangerous word, the process proceeds to step 2004 without performing the process of step 2003.

[0033] In step 2004, it is determined whether the call has ended. If the call has not ended (NO in step 2004), the process returns to step 2002 again, and a process of determining whether or not there is a dangerous word in the voice-recognized text is performed. If it is determined in step 2004 that the call has ended (YES in step 2004), the danger is determined according to the level of the danger level X, and the danger level determination process ends.

[0034] FIG. 7 is a table showing an example of the relationship between dangerous words and their risk levels, as an example of dangerous word information stored in memory unit 13 (see FIG. 4). In FIG. 7, frequently occurring phrases in specialized frauds are listed in the left column of the table, and risk level values ​​are listed in the right column, which are associated with each other and stored in memory unit 13. For example, if a frequently occurring word in specialized frauds is "refund," the risk level value is "2." Also, for example, the risk level value for the phrase "refund coming back" is "3."

[0035] The relationship between the dangerous words and the danger level as exemplified in Fig. 7 is set, for example, by the server 30. For phrases or keywords that appear during a call, a danger level value is set according to the likelihood of a special fraud. The danger level may be set by machine learning using artificial intelligence, or the administrator of the server may set the danger level.

[0036] In step 2003 in FIG. 6-2 described above, the risk level according to the dangerous word is added to the risk level X using the information exemplified in FIG. 7. In addition to determining the risk according to the level of risk level X as shown in step 2005, the information to be transmitted may be changed according to the level of risk level X. For example, when the level of risk level X is low, the tone may be lowered to persuade the user, and when the level of risk level X is high, the tone may be raised to persuade the user. In addition, when the level of risk level X is high, it is also effective to change the notification destination or notification content, such as contacting the police.

[0037] Next, the emotion recognition process in step 1004 shown in FIG. 6-1 will be described with reference to FIG. 6-3. FIG. 6-3 is a flowchart showing a process for recognizing a user's emotion. First, the emotion recognition unit 114 acquires user feature points from the information on user characteristics stored in the storage unit 13 (step 3001). Next, the user characteristic recognition unit 115 extracts feature points from information on the user during a call that is included in the call information (step 3002). Then, the emotion recognition unit 114 compares the feature points of the user characteristics with the feature points of the call information (step 3003). After that, the emotion recognition unit 114 recognizes the user's emotion based on the comparison (step 3004), and the emotion recognition process ends.

[0038] Next, an example in which the persuasive information generating unit 116 (see FIG. 4) of the communication robot 10 generates persuasive information will be described. In general, once elderly people or other victims of fraud believe the words of a fraudster, they tend to stubbornly believe the words of the fraudster as true, even if people around them directly deny that it is a fraud. When a person is caught up in a preconceived notion like this, it is effective to have them recall events or memories that they found enjoyable, and to distract them from the things they are thinking about. Therefore, persuasive information is generated for users who receive calls about fraudulent activities, in order to evoke events or memories that the user found enjoyable.

[0039] The user characteristic recognition unit 115 (see FIG. 4) recognizes events that the user finds enjoyable through daily conversations between the user and the communication robot 10, and stores them in the storage unit 13 as user characteristic information. For example, when an event occurs that results in lively conversation with the user, the event is stored as "user's favorite object." Examples include the user's favorite person, a celebrity the user likes, the user's hobbies, a sport that the user frequently watches, an event the user participates in, the user's favorite food, and information about a store. For example, an example of a user's favorite person may include information about a grandchild.

[0040] When generating persuasive information, the persuasive information generating unit 116 reads out the user characteristic information stored in the storage unit 13 and uses it to generate the persuasive information. For example, depending on the persuasive situation, one of the "user's favorite targets" is selected. Then, the selected target is used to generate persuasive information. An example of the generated persuasive information is, for example, for a user who loves "Ai-chan" of "grandchild", "Ai-chan came over last week and it was fun. I wonder how Ai-chan is doing? While you're listening to Ai-chan's voice, why don't you talk to Ai-chan's mother about it?". Here, persuasive information is generated that recalls memories that the user himself felt were fun, distracts the user from matters that he believes, and goes as far as to suggest measures to resolve the problem. For a user who usually talks about baseball with the communication robot 10, persuasive information is generated that directly stimulates the user's memories that he felt were fun, such as "Yesterday's baseball game was amazing. Which scene was good?". These persuasive information arouses a comfortable feeling in a user who is at high risk of fraud, and gives the user an opportunity to distract the user from the words of the fraudster.

[0041] As described above, after distracting the person who is obsessed with a preconceived notion from the matter he / she is preconceived, the persuasive information generating unit 116 generates specific persuasive information. Here, an example of generating persuasive information using "user's personality" as a characteristic of the user will be described. For example, for a user whose personality is "calm," persuasive information is generated using a dialogue scenario created for a person with a "calm" personality. An example of persuasive information generated using a dialogue scenario for a person with a "calm" personality is, "According to statistical data, when cold, voice, and money come up during a conversation, there is a 92% chance that it is an "it's me" scam." For a user whose personality is "calm," persuasive information is generated that presents the reasons. Examples of persuasive information generated for a user with a "stubborn" personality include "That may be an "it's me" scam" and "You should contact your family." For a user with a "stubborn" personality, persuasive information with ambiguous expressions is generated, rather than definitive expressions such as "That's an "it's me" scam" and "Contact your family." Examples of persuasive information generated for a user with a "cooperative" personality include "In a situation like this, most people would call their family to check," or "In this case, many people would call the police." For a user with a "cooperative" personality, persuasive information is generated by presenting information about how other people generally respond. These persuasion examples are merely examples, but it is sufficient that desirable persuasion information is generated through daily dialogue between the communication robot 10 and the user.

[0042] Although the embodiment of the present invention has been described above, the technical scope of the present invention is not limited to the scope described in the above embodiment. It is clear from the claims that various modifications and improvements to the above embodiment are also included in the technical scope of the present invention. [Explanation of symbols]

[0043] 1...robot warning system, 10...communication robot, 30...server, 40...database, 50...network, 60...connection device, 70...telephone, 80...terminal owned by user's related party, 111...call information processing unit, 112...communication information input unit, 113...danger level determination unit, 114...emotion recognition unit, 115...user characteristic recognition unit, 116...persuasion information generation unit, 117...communication information output unit

Claims

1. An interaction means for interacting with a user; A call information acquisition means for acquiring information on a call between the user and a third party; a risk determination means for determining a degree of risk that the call is fraudulent using the acquired call information; A characteristic acquisition means for acquiring information regarding individual characteristics of the user; An emotion information acquisition means for acquiring information regarding the emotion of the user during or after a call; Equipped with the dialogue means, when it is determined that the degree of risk is high, determines and outputs dialogue contents based on the information on the individual characteristics and the information on the emotion; the information on the emotion is information indicating a degree to which the user is upset, the information being determined based on at least one of voice information of the user and a facial expression of the user included in the call information; The information indicating the degree of agitation is information obtained by comparing history information regarding a dialogue previously conducted between the dialogue means and the user and information regarding the emotion acquired by the emotion information acquisition means. A communication robot that features:

2. The information about each characteristic is a characteristic behavior pattern of the user that is understood from the history information. The communication robot according to claim 1 .

3. the dialogue means changes at least one of the contents of the voice and the characteristics of the voice from the stored voice information in accordance with the individual characteristic behavioral pattern of the user, or dialogues with the user using new voice information; The communication robot according to claim 2 .

4. The information about each characteristic is information about an event that has generated a lively dialogue with the user; 2. The communication robot according to claim 1, wherein the dialogue means determines the content of the dialogue based on information regarding an event that has led to a lively dialogue with the user.

Images