JPWO2021250778A5 - - Google Patents

Description

The present invention relates to a conversation control program, a conversation control method, and an information processing apparatus.

In recent years, an increasing number of companies are introducing automatic response systems that use technology called chatbots. Chatbots are computer systems that use artificial intelligence to interact and exchange messages with humans. For example, by utilizing two-way communication in cooperation with a chatbot, there is a case where a chatbot is used to perform window services that have conventionally been handled by people.

As a prior art, for example, when chat data transmitted from a user terminal is transferred to a chatbot via a network and it is detected that the chat data transmitted from the user terminal contains content related to personal information, There is a chat system that performs predetermined filtering on the chat data.

Japanese Unexamined Patent Application Publication No. 2018-200602

However, in the conventional technology, when performing a service using two-way communication linked with a chatbot, it is necessary to input information that can identify an individual, which may hinder the use of the service using the chatbot.

In one aspect, an object of the present invention is to facilitate smooth use of services using chatbots.

In one embodiment, input information accepted in a chatbot conversation is received from an information processing terminal, response processing is executed based on the input information, and when a request for masking processing of the input information is received, the chat Provided is a conversation control program for executing a masking process for not recording or concealing a log regarding input information among logs regarding a conversation when recording a log regarding a conversation with a bot.

ADVANTAGE OF THE INVENTION According to the one aspect of this invention, it is effective in being able to aim at smooth utilization of the service using a chatbot.

FIG. 1 is an explanatory diagram showing a system configuration example of a conversation control system 100 according to the first embodiment. FIG. 2 is a block diagram showing a hardware configuration example of the information processing apparatus 101. As shown in FIG. FIG. 3 is a block diagram showing a hardware configuration example of the information processing terminal 102. As shown in FIG. FIG. 4 is an explanatory diagram showing an example of the contents of the script table 130. As shown in FIG. FIG. 5 is an explanatory diagram showing an example of the contents stored in the chat log DB 140. As shown in FIG. FIG. 6 is an explanatory diagram showing a specific example of a script. FIG. 7 is an explanatory diagram (Part 1) showing an example of a chat screen. FIG. 8 is a block diagram of a functional configuration example of the information processing apparatus 101 according to the first embodiment. FIG. 9 is an explanatory diagram (part 2) showing an example of a chat screen. FIG. 10 is an explanatory diagram showing an operation example of the information processing apparatus 101. As shown in FIG. FIG. 11 is a flowchart (part 1) showing an example of a conversation control processing procedure of the information processing apparatus 101 according to the first embodiment. FIG. 12 is a flowchart (part 2) showing an example of the conversation control processing procedure of the information processing apparatus 101 according to the first embodiment. FIG. 13 is a block diagram of a functional configuration example of the information processing apparatus 101 according to the second embodiment. FIG. 14 is an explanatory diagram showing an example of the contents stored in the number counter table 1400. As shown in FIG. FIG. 15 is a flowchart (part 1) showing an example of the first conversation control processing procedure of the information processing apparatus 101 according to the second embodiment. FIG. 16 is a flowchart (part 2) showing an example of the first conversation control processing procedure of the information processing apparatus 101 according to the second embodiment. FIG. 17 is a flowchart (part 1) showing an example of the second conversation control processing procedure of the information processing apparatus 101 according to the second embodiment. FIG. 18 is a flowchart (part 2) showing an example of the second conversation control processing procedure of the information processing apparatus 101 according to the second embodiment. FIG. 19 is a block diagram of a functional configuration example of the information processing apparatus 101 according to the third embodiment. FIG. 20 is an explanatory diagram showing an example of the contents stored in the personal identification information master 2000. As shown in FIG. FIG. 21 is an explanatory diagram showing a specific example of learning data. FIG. 22 is an explanatory diagram of an example of password determination. FIG. 23 is an explanatory diagram (part 3) showing an example of a chat screen. FIG. 24 is an explanatory diagram (part 4) showing an example of a chat screen. FIG. 25 is a flowchart (part 1) showing an example of the first conversation control processing procedure of the information processing apparatus 101 according to the third embodiment. FIG. 26 is a flowchart (part 2) showing an example of the first conversation control processing procedure of the information processing apparatus 101 according to the third embodiment. FIG. 27 is a flowchart showing an example of the second conversation control processing procedure of the information processing apparatus 101 according to the third embodiment.

Exemplary embodiments of a conversation control program, conversation control method, and information processing apparatus according to the present invention will be described below in detail with reference to the drawings.

(Embodiment 1)
First, a system configuration example of the conversation control system 100 including the information processing device 101 according to the first embodiment will be described. The conversation control system 100 is applied, for example, to a computer system that provides services using two-way communication linked with a chatbot.

FIG. 1 is an explanatory diagram showing a system configuration example of a conversation control system 100 according to the first embodiment. In FIG. 1 , a conversation control system 100 includes an information processing device 101 , a plurality of information processing terminals 102 (three in the example of FIG. 1), and an administrator terminal 103 . In conversation control system 100 , information processing device 101 , information processing terminal 102 and administrator terminal 103 are connected via wired or wireless network 110 . The network 110 is, for example, a LAN (Local Area Network), a WAN (Wide Area Network), the Internet, or the like.

Here, the information processing device 101 is a computer that controls conversation of a chatbot. The information processing device 101 is, for example, a server. The information processing apparatus 101 has, for example, a FAQ master 120, a script table 130, and a chat log DB (Database) 140. FIG. The FAQ master 120 stores FAQs (Frequently Asked Questions).

FAQ means "Frequently Asked Questions", and is a collection of questions and answers to expected questions. The question may ask how to solve some problem, or it may be a spoken question. An answer may be an answer that indicates how to solve the question asked, or an answer to a spoken question.

The script table 130 stores scripts that define the conversation flow of the chatbot. Chatbot conversations are scripted. The chat log DB 140 stores chat logs. A chat log is a history of conversations between a user and a chatbot.

Note that the storage contents of the script table 130 and the chat log DB 140 will be described later with reference to FIGS. 4 and 5. FIG.

The information processing terminal 102 is a computer that a user uses to have a chatbot conversation. The information processing terminal 102 is, for example, a smart phone, a tablet PC (Personal Computer), a PC, or the like. Also, the information processing terminal 102 may be a digital board installed at stations, airports, shops, commercial facilities, and the like.

In the information processing terminal 102, the user can, for example, enter a message or select an option to search for a desired answer or perform an online procedure while interacting with the chatbot. An online procedure is a procedure that is carried out online. Online procedures include, for example, procedures such as account opening, balance inquiry, and transfer provided by financial institutions, and procedures such as communication service plan change and model change provided by communication carriers.

An administrator terminal 103 is a computer used by an administrator of the conversation control system 100 . The administrator of the conversation control system 100 is, for example, the designer of the scripts that define the conversation flow of the chatbot. The administrator terminal 103 is, for example, a PC. At the administrator terminal 103, the administrator can, for example, design and change scripts and set various thresholds.

Although illustration is omitted, the conversation control system 100 may include an operator terminal. An operator terminal is a computer used by an operator. The operator is a person who responds to the user when the conversation is switched from chatbot support to manned support. For example, the operator can communicate with the user using the telephone function of the operator terminal, text communication, and the like.

Here, in a service that utilizes two-way communication linked with a chatbot, there are cases where the input of information that can identify an individual is required. For example, in some consultations, the input of information that can identify an individual such as name and address may be requested. In this case, it becomes possible to specify who consulted what kind.

Users with high IT literacy know that the information they enter is recorded as a log on the server, and they are aware that such information may be used for marketing and product recommendations. many. A user who dislikes the use of his/her own information may stop using the service when requested to input information.

Note that it is conceivable not to record in the log any information that the user may not want to know. However, the reality is that business operators want to receive and utilize data, and it is not desirable to uniformly prevent information that users may not want to know from being recorded.

Therefore, in the first embodiment, when the information processing apparatus 101 receives input information received in a chatbot conversation, it executes response processing based on the input information. When recording a log related to a chatbot conversation, a masking process is executed for not recording or anonymizing the log related to the input information among the logs related to the conversation. A conversation control method for smooth use of services using chatbots by allowing users to control recording in logs by themselves so that users can use services with peace of mind will be described.

(Hardware Configuration Example of Information Processing Device 101)
Next, a hardware configuration example of the information processing apparatus 101 will be described.

FIG. 2 is a block diagram showing a hardware configuration example of the information processing apparatus 101. As shown in FIG. 2, an information processing apparatus 101 includes a CPU (Central Processing Unit) 201, a memory 202, a disk drive 203, a disk 204, a communication I/F (Interface) 205, and a portable recording medium I/F 206. , and a portable recording medium 207 . Also, each component is connected by a bus 200 .

Here, the CPU 201 controls the entire information processing apparatus 101 . The CPU 201 may have multiple cores. The memory 202 has, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), a flash ROM, and the like. Specifically, for example, a flash ROM stores an OS (Operating System) program, a ROM stores an application program, and a RAM is used as a work area for the CPU 201 . A program stored in the memory 202 is loaded into the CPU 201 to cause the CPU 201 to execute coded processing.

The disk drive 203 controls data read/write with respect to the disk 204 under the control of the CPU 201 . The disk 204 stores data written under the control of the disk drive 203 . Examples of the disk 204 include a magnetic disk and an optical disk.

The communication I/F 205 is connected to the network 110 through a communication line, and is connected to an external computer (for example, the information processing terminal 102 and administrator terminal 103 shown in FIG. 1) via the network 110 . A communication I/F 205 serves as an interface between the network 110 and the inside of the apparatus, and controls input/output of data from an external computer. For the communication I/F 205, for example, a modem, a LAN adapter, or the like can be adopted.

A portable recording medium I/F 206 controls reading/writing of data from/to a portable recording medium 207 under the control of the CPU 201 . The portable recording medium 207 stores data written under control of the portable recording medium I/F 206 . Examples of the portable recording medium 207 include CD (Compact Disc)-ROM, DVD (Digital Versatile Disk), USB (Universal Serial Bus) memory, and the like.

Note that the information processing apparatus 101 may include, for example, an SSD (Solid State Drive), an input device, a display, etc., in addition to the components described above. Further, the information processing apparatus 101 does not have to include, for example, the disk drive 203, the disk 204, the portable recording medium I/F 206, and the portable recording medium 207 among the components described above.

(Hardware configuration example of information processing terminal 102)
Next, a hardware configuration example of the information processing terminal 102 will be described.

FIG. 3 is a block diagram showing a hardware configuration example of the information processing terminal 102. As shown in FIG. 3, information processing terminal 102 includes CPU 301, memory 302, display 303, input device 304, communication I/F 305, camera 306, speaker 307, microphone 308, and GPS (Global Positioning System). a unit 309; Also, each component is connected by a bus 300 .

Here, the CPU 301 controls the entire information processing terminal 102 . The CPU 301 may have multiple cores. The memory 302 is a storage unit having, for example, ROM, RAM and flash ROM. Specifically, for example, a flash ROM or ROM stores various programs, and a RAM is used as a work area for the CPU 301 . A program stored in the memory 302 is loaded into the CPU 301 to cause the CPU 301 to execute coded processing.

A display 303 is a display device that displays data such as a cursor, an icon, or a toolbox, as well as documents, images, and functional information. As the display 303, for example, a liquid crystal display, an organic EL (Electroluminescence) display, or the like can be adopted.

The input device 304 has keys for inputting characters, numbers, various instructions, etc., and inputs data. The input device 304 may be a touch panel type input pad or numeric keypad, or may be a keyboard or mouse.

The communication I/F 305 is connected to the network 110 through a communication line, and is connected to an external computer (for example, the information processing apparatus 101) via the network 110. FIG. A communication I/F 305 serves as an interface between the network 110 and the inside of the device, and controls input/output of data from an external device.

A camera 306 is a photographing device that photographs an image (still image or moving image) and outputs image data. The speaker 307 converts the electrical signal into sound and outputs the sound. A microphone 308 receives sound and converts it into an electrical signal.

The GPS unit 309 receives radio waves from GPS satellites and outputs position information of the information processing terminal 102 . The position information of the information processing terminal 102 is, for example, information specifying one point on the earth such as latitude and longitude. Also, as the satellite, for example, a satellite of the quasi-zenith satellite system may be used.

In addition to the components described above, the information processing terminal 102 includes, for example, an HDD (Hard Disk Drive), an SSD, a short-range wireless communication I/F, a portable recording medium I/F, a portable recording medium, and the like. You can decide. Also, the administrator terminal 103 shown in FIG. 1 can be realized by a hardware configuration similar to that of the information processing terminal 102 .

(Stored Contents of Script Table 130)
Next, contents stored in the script table 130 of the information processing apparatus 101 will be described with reference to FIG. The script table 130 is realized by a storage device such as the memory 202 and the disk 204 shown in FIG. 2, for example.

FIG. 4 is an explanatory diagram showing an example of the contents of the script table 130. As shown in FIG. In FIG. 4, the script table 130 has service and script fields, and sets information in each field to store script management information 400-1 to 400-n as records.

Here, the service indicates a service using two-way communication linked with a chatbot. Here, only procedure IDs (for example, A1 and A2) that identify services are shown. As a service, for example, there is a service that uses two-way communication linked with a chatbot to provide consultation on health, further education, career advancement, etc., and to perform various online procedures.

A script is information (scenario) that defines the flow of chatbot conversation, and is prepared in association with a service. However, only script IDs (for example, S1 and S2) for identifying scripts are shown here. For example, the script management information 400-1 indicates script S1 corresponding to service A1.

(Stored content of chat log DB 140)
Next, contents stored in the chat log DB 140 of the information processing apparatus 101 will be described with reference to FIG. The chat log DB 140 is implemented, for example, by a storage device such as the memory 202 and disk 204 shown in FIG.

FIG. 5 is an explanatory diagram showing an example of the contents stored in the chat log DB 140. As shown in FIG. In FIG. 5, the chat log DB 140 has chat_id, talklog_id, time stamp, script, label and message fields. By setting information in each field, chat logs (for example, chat logs 500-1 to 500-5) are stored as records.

Here, chat_id indicates the ID of the chat room. chat_id can identify a series of conversations in a chatbot. talklog_id is an identifier that identifies a chat log. The time stamp indicates the date and time (GMT+9 hours) when the talk process (step) defined in the script was executed. GMT is Greenwich Mean Time.

The talk process is a process for realizing the operation (behavior) of the chatbot. For example, the talk process includes a process of speaking, a process of displaying options, a process of selecting an option according to user operation, a process of performing a procedure according to the selected option or input information, and a process of ending the talk process. and so on.

script indicates a script (scenario) applied to the chatbot conversation. label indicates the name of the step in which talk processing is performed among the steps in the script. The message indicates a sentence uttered in the executed talk process, a selected option, input information, processing contents on the system side, an error message, or the like.

For example, chat log 500-2 is a log with chat_id “100”, talklog_id “15044” and time stamp “2020-01-08T14:25:43+09:00”. According to the chat log 500-2, it is specified that the talk processing (step) of label "VRB-1" of script "S10" was executed and the message "What is your name?" was displayed. can be done.

The chat log may include, for example, information indicating whether or not it is a system log, and information indicating whether the log relates to a chatbot or a user. The chat log may also include, for example, the IP (Internet Protocol) address of the information processing terminal 102 used by the user.

A specific example of the script will now be described with reference to FIG. Here, the script S100 corresponding to the service A100 will be described as an example.

FIG. 6 is an explanatory diagram showing a specific example of a script. In FIG. 6, a script S100 is information (scenario) that defines the conversation flow of the chatbot. However, in FIG. 6, a part of the script S100 is extracted and displayed. Specifically, for example, the script S100 is information that defines chat processing.

Script S100 includes steps 601-605. Here, step 601 indicates the start of talk processing. Steps 602 and 603 show the process of speaking a message. Step 604 shows the process of displaying options. Step 605 indicates the end of the talk process.

Note that the same label (see FIG. 5) may be set for a plurality of steps. For example, steps 602 and 603 are set to the same "VRB-1". Although not shown, the script S100 includes, in addition to steps 601 to 605, steps for executing procedures and FAQ searches according to the selected options and input information. be

(Screen example of chat screen)
Next, an example of a chat screen displayed on the display 303 (FIG. 3) of the information processing terminal 102 will be described.

FIG. 7 is an explanatory diagram (Part 1) showing an example of a chat screen. In FIG. 7, a chat screen 700 is an example of an operation screen that displays a chatbot conversation on the information processing terminal 102 . In FIG. 7, avatar av is a character representing a chatbot.

On the chat screen 700, for example, a message can be input by selecting a box 710 through user operation input using the input device 304 shown in FIG. Input of the message may be performed by voice input using the microphone 308 shown in FIG. 3, for example. Also, by selecting a send button 711 on the chat screen 700, the message input in the box 710 can be sent to the chatbot side (information processing apparatus 101).

On the chat screen 700, a message m1 representing the statement of the avatar av is displayed. Message m1 is an example of a message prompting the user to enter the name. Also, on the chat screen 700, a message m2 representing the user's statement is displayed. Message m2 is an example of the name entered for message m1 of avatar av.

A secret button 720 is also displayed on the chat screen 700 . The secret button 720 is an example of information for inquiring whether or not to permit recording of input information in the log. The input information is information that indicates input information and selected options in a chatbot conversation.

On the chat screen 700, for example, when the secret button 720 is selected by the user's operation input, the off-the-record mode can be set. The off-the-record mode is a mode in which input information is not recorded in the log or is made anonymous when recording a log (chat log) relating to a chatbot conversation.

As a result, the user himself/herself can control information that is not recorded in the log or that is made anonymous among the information that is input in the chatbot conversation.

(Example of functional configuration of information processing apparatus 101)
FIG. 8 is a block diagram of a functional configuration example of the information processing apparatus 101 according to the first embodiment. In FIG. 8 , the information processing apparatus 101 includes a reception unit 801 , a conversation control unit 802 and a recording control unit 803 . Specifically, for example, the reception unit 801 to the recording control unit 803 cause the CPU 201 to execute a program stored in a storage device such as the memory 202, the disk 204, and the portable recording medium 207 shown in FIG. Alternatively, the function is realized by the communication I/F 205 . The processing results of each functional unit are stored in a storage device such as the memory 202 or disk 204, for example.

The reception unit 801 receives a chat start instruction from the information processing terminal 102 . Here, the chat start instruction is for starting a conversation between the user and the chatbot, for example, requesting the start of a service using the chatbot. The chat start instruction includes, for example, a service ID.

For example, an instruction to start a chat is input by a user at the information processing terminal 102 and transmitted from the information processing terminal 102 to the information processing apparatus 101 . Specifically, for example, the accepting unit 801 accepts the chat start instruction by receiving the chat start instruction from the information processing terminal 102 .

The conversation control unit 802 starts a chatbot conversation on the information processing terminal 102 in response to a chat start instruction from the information processing terminal 102 . Specifically, for example, the conversation control unit 802 refers to the script table 130 shown in FIG. 4 to identify the script corresponding to the service ID included in the chat start instruction.

Next, the conversation control unit 802 determines the identified script as a talk script to be applied to the conversation of the chatbot on the information processing terminal 102 . Then, the conversation control unit 802 starts a chat regarding the service on the information processing terminal 102 using the decided talk script.

For example, assume that the service ID "A1" is included in the chat start instruction. In this case, the conversation control unit 802 refers to the script table 130 and determines the script S1 corresponding to the service ID "A1" as the talk script. Then, the conversation control unit 802 uses the determined talk script S1 to start a chatbot conversation regarding the service S1 in the information processing terminal 102 .

Note that the conversation control unit 802 may determine a predetermined script as a talk script in response to a chat start instruction from the information processing terminal 102 . Also, the conversation control unit 802 may determine a talk script based on, for example, the user attribute of the information processing terminal 102 or the external environment. A user attribute is information representing characteristics of a user of the information processing terminal 102 . User attributes represent, for example, age, gender, language used, and the like. The external environment is information representing the external environment of the information processing terminal 102 . The external environment represents, for example, weather, temperature, humidity, time, location, and the like.

When the conversation control unit 802 receives input information accepted in the chatbot conversation from the information processing terminal 102, the conversation control unit 802 executes response processing based on the input information. Here, the input information is information input by a user's operation using the input device 304 shown in FIG. 3, for example, in a chatbot conversation. The input information may also be information indicating an option selected by a user's operation in a chatbot conversation.

The response process based on the input information is, for example, a process of searching the FAQ according to the information input in the chatbot conversation and outputting the result. The FAQ search is performed using, for example, the FAQ master 120 shown in FIG. Also, the response processing based on the input information may be processing for performing a procedure according to information input in a chatbot conversation.

As a result of the conversation control unit 802 starting a chatbot conversation on the information processing terminal 102 , the recording control unit 803 records a log of the chatbot conversation. Specifically, for example, the recording control unit 803 records a chat log when a user or a chatbot makes a statement or when an internal process of the system is performed. The input information received in the chatbot conversation is included in, for example, the message of the chat log (see FIG. 5).

Further, when the recording control unit 803 receives from the information processing terminal 102 a masking process request for input information accepted in the chatbot conversation, the recording control unit 803 performs masking of the input information when recording a log of the chatbot conversation. Execute the process. Here, the masking processing request is for requesting execution of the masking processing.

The masking process is a process of not recording or anonymizing logs related to input information among logs related to chatbot conversations. Not recording in a log may mean, for example, creating a chat log in which input information is not set in a message and the message is empty. Also, not recording in the log may mean deleting the input information from the chat log after once creating a chat log in which the input information is set in the message.

Anonymization means, for example, converting input information into information that cannot be guessed. Information that cannot be inferred is, for example, an asterisk (*) or blackout. Anonymization may mean converting input information into information that cannot be inferred before setting it in a message, or converting input information set in a message into information that cannot be inferred. .

Note that the information processing apparatus 101 may receive input information and a masking processing request at the same time, for example, in a chatbot conversation. Further, the information processing apparatus 101 may receive a masking processing request for the input information after receiving the input information. Further, the information processing apparatus 101 may receive input information to be subjected to masking processing after receiving the masking processing request.

Specifically, for example, the conversation control unit 802 outputs, to the information processing terminal 102, inquiry information as to whether or not to permit recording of input information in a log in a chatbot conversation. The masking processing request is transmitted from the information processing terminal 102, for example, when an operation is performed on the information processing terminal 102 to the effect that recording to the log is not permitted in response to the inquiry information.

The inquiry information is a message, an icon, or the like for inquiring whether or not to permit recording of the input information in the log, and is, for example, the secret button 720 shown in FIG. More specifically, for example, when starting a chatbot conversation on the information processing terminal 102 , the information processing terminal 102 is caused to display a chat screen 700 including a secret button 720 .

When the secret button 720 is selected by the user's operation input on the chat screen 700 , an off-the-record setting instruction is transmitted from the information processing terminal 102 to the information processing apparatus 101 . The off-record setting instruction corresponds to a masking processing request. Upon receiving the off-record setting instruction from the information processing terminal 102, the recording control unit 803 sets (turns on) the off-record mode.

The off-the-record mode is a mode in which input information is not recorded in the log or is made anonymous when recording a log (chat log) relating to a chatbot conversation. That is, in the off-the-record mode, the user's remarks are not recorded in the log as information to be masked, or are recorded in the log in an anonymized manner.

Specifically, for example, the recording control unit 803 determines whether or not the off-the-record mode is set when input information accepted in a chatbot conversation is received from the information processing terminal 102 . Here, if the off-the-record mode is not set, the recording control unit 803 does not perform masking processing on the received input information when recording the chat log.

On the other hand, when the off-the-record mode is set, the recording control unit 803 performs masking processing on the received input information when recording the chat log. More specifically, for example, the recording control unit 803 may delete the input information from the chat logs related to the received input information among the chat logs related to the chatbot conversation. The input information in the chat log may be converted into information that cannot be guessed, such as asterisks or blackouts, to be made anonymous.

Although the case where the secret button 720 is pressed before inputting information to be masked has been described as an example, the present invention is not limited to this. For example, the secret button 720 may be pressed after inputting information to be masked. In this case, the recording control unit 803 performs masking processing on input information received after receiving an off-record setting instruction (masking processing request).

Further, it may be possible to designate a part of the input information received in the chatbot conversation as information to be subjected to masking processing. For example, on the chat screen 700, it may be possible to designate part of the information in the message input to the box 710 as the masking target information by the user's operation input using the input device 304. FIG. In this case, for example, the recording control unit 803 may execute masking processing on information (for example, part of an address) specified as information to be masked among the received input information. Information other than the designated information among the input information is recorded in the log as it is, for example.

Also, the chat log is recorded in the chat log DB 140 shown in FIG. Specifically, for example, a chat log recorded during a chatbot conversation is temporarily stored in a buffer within the memory 202, for example. When the chatbot conversation ends, all chat logs in the buffer are output to the chat log DB 140 . A chat log 500-1 shown in FIG. 5 is an example of a log including masked input information.

Further, for example, on the chat screen 700, information that has already been input as information not subject to masking processing may be additionally designated as information subject to masking processing before the conversation in the chatbot ends. In this case, the recording control unit 803 performs, for example, a masking process that deletes or anonymizes the information from the additionally specified information to be masked, among the logs temporarily stored in the buffer in the memory 202 . Execute the process.

Each functional unit of the information processing device 101 described above may be realized by a plurality of computers (for example, the information processing device 101 and the information processing terminal 102) in the conversation control system 100. FIG. Also, each functional unit of the information processing apparatus 101 described above may be realized by the information processing terminal 102 .

(Example of operation of information processing device 101)
Next, an operation example of the information processing apparatus 101 according to the first embodiment will be described with reference to FIGS. 9 and 10. FIG. Here, a chat screen that is displayed when health consultation is performed using two-way communication linked with a chatbot will be described as an example.

FIG. 9 is an explanatory diagram (part 2) showing an example of a chat screen. FIG. 10 is an explanatory diagram showing an operation example of the information processing apparatus 101. As shown in FIG. In FIG. 9, the chat screen 700 displays a message m3 representing the utterance of the avatar av. The message m3 is an example of a message that prompts the user to input consultation content.

Here, the user's name "Fuji Taro" is input prior to inputting the user's consultation content. The name "Fuji Taro" is information that can identify an individual. Therefore, when the content of the user's consultation is input in response to the message m3, information that can identify who consulted and what kind of consultation is recorded in the log.

Here, it is assumed that the user selects the secret button 720 on the chat screen 700 because the user has determined that he/she does not want to record the content of his or her consultation in the log. In this case, an off-record setting instruction is transmitted from the information processing terminal 102 to the information processing apparatus 101, and the off-record mode is set (ON).

When the secret button 720 is selected, for example, the display mode of the secret button 720 is changed so that it can be determined that the off-record mode is being set (see FIG. 10, for example). Also, when the secret button 720 is selected, for example, an icon indicating that the off-the-record mode is being set may be displayed.

Further, when the secret button 720 is selected again on the chat screen 700 and then the secret button 720 is selected again, for example, an instruction to cancel the off-record mode is transmitted from the information processing terminal 102 to the information processing apparatus 101, and the off-record mode is canceled ( OFF). Also, the off-the-record mode may be automatically canceled each time input information is transmitted from the information processing terminal 102 to the information processing apparatus 101 .

In FIG. 10, a chat screen 700 displays a message m4 representing the user's statement. The message m4 is an example of the content of the consultation input for the message m3 of the avatar av. The input message m4 is transmitted from the information processing terminal 102 to the information processing apparatus.

For example, when receiving the message m4, the information processing apparatus 101 refers to the FAQ master 120 and searches FAQ based on the message m4. The information processing apparatus 101 then outputs the search result to the information processing terminal 102 . As a result, on the chat screen 700, for example, the searched result (not shown) is displayed in a selectable manner as the utterance of the avatar av.

Thereby, the user can obtain an answer (FAQ) to the content of consultation (question) input by the user.

Also, here, the off-record mode is set. Therefore, the information processing apparatus 101 performs masking processing on the message m4 when recording a chat log regarding the message m4 received in the chatbot conversation.

Specifically, for example, the information processing apparatus 101 records a chat log 500-5 in which the received message m4 is deleted. The message "***deleted***" in the chat log 500-5 indicates that the input information has been deleted. In addition, in FIG. 10, a part of the chat log is extracted and displayed.

(Conversation control processing procedure of information processing device 101)
Next, a conversation control processing procedure of the information processing apparatus 101 according to the first embodiment will be described with reference to FIGS. 11 and 12. FIG.

11 and 12 are flowcharts showing an example of the conversation control processing procedure of the information processing apparatus 101 according to the first embodiment. In the flowchart of FIG. 11, first, the information processing apparatus 101 determines whether or not a chat start instruction has been received from the information processing terminal 102 (step S1101).

Here, the information processing apparatus 101 waits to receive a chat start instruction (step S1101: No). When the information processing apparatus 101 receives the chat start instruction (step S1101: Yes), the information processing apparatus 101 refers to the script table 130 and determines the script corresponding to the service ID included in the received chat start instruction as the talk script. (step S1102).

Then, the information processing apparatus 101 uses the determined talk script to start a chatbot conversation on the information processing terminal 102 (step S1103). At this time, the information processing apparatus 101 outputs to the information processing terminal 102 inquiry information as to whether or not to permit recording of the input information in the log. The inquiry information is, for example, the confidential button 720 shown in FIG.

The information processing apparatus 101 then determines whether or not an off-record setting instruction has been received from the information processing terminal 102 (step S1104). Here, when the off-record setting instruction is received from the information processing terminal 102 (step S1104: Yes), the information processing apparatus 101 turns on the off-record mode (step S1105), and proceeds to step S1201 shown in FIG.

On the other hand, if the off-the-record setting instruction has not been received from the information processing terminal 102 (step S1104: No), the information processing apparatus 101 determines whether or not the off-the-record cancellation instruction has been received from the information processing terminal 102 (step S1106). .

Here, if the off-the-record cancellation instruction has not been received from the information processing terminal 102 (step S1106: No), the information processing apparatus 101 proceeds to step S1201 shown in FIG. On the other hand, when the off-record canceling instruction is received from the information processing terminal 102 (step S1106: Yes), the information processing apparatus 101 turns off the off-record mode (step S1107), and proceeds to step S1201 shown in FIG.

In the flowchart of FIG. 12, first, the information processing apparatus 101 determines whether or not input information accepted in the chatbot conversation is received from the information processing terminal 102 (step S1201). If the input information has not been received (step S1201: No), the information processing apparatus 101 proceeds to step S1206.

On the other hand, when input information is received (step S1201: Yes), the information processing apparatus 101 executes response processing based on the received input information (step S1202). The information processing apparatus 101 then determines whether the off-record mode is ON (step S1203).

If the off-the-record mode is ON (step S1203: Yes), the information processing apparatus 101 performs masking processing on the received input information and outputs a chat log in which the input information is masked (step S1204). The process moves to step S1206.

On the other hand, if the off-record mode is OFF (step S1203: No), the information processing apparatus 101 outputs a chat log including the received input information (step S1205). Then, the information processing apparatus 101 determines whether or not to end the chatbot conversation on the information processing terminal 102 (step S1206).

Here, if the conversation is not to be ended (step S1206: No), the information processing apparatus 101 returns to step S1104 shown in FIG. On the other hand, when ending the conversation (step S1206: Yes), the information processing apparatus 101 ends the series of processes according to this flowchart. The chatbot conversation ends, for example, according to the talk script or according to the user's forced termination operation.

As a result, in response to an off-the-record setting instruction from the information processing terminal 102, it is possible to mask the input information received in the chatbot conversation and output it as a log.

As described above, according to the information processing apparatus 101 according to the first embodiment, it is possible to receive input information accepted in a chatbot conversation from the information processing terminal 102 and execute response processing based on the input information. can. Further, according to the information processing apparatus 101, when a request for masking input information is received, when recording a log regarding a chatbot conversation, the log regarding the input information among the logs regarding the conversation is not recorded or is kept confidential. A masking process that transforms the

As a result, in response to the masking processing request from the user, response processing (for example, FAQ search, etc.) based on the input information received in the chatbot conversation is executed, and the input information is not left in the log. can do.

Further, according to the information processing apparatus 101, in a chatbot conversation, it is possible to output to the information processing terminal 102 inquiry information as to whether or not to permit recording of input information in a log. Then, according to the information processing apparatus 101, when an operation is performed on the information processing terminal 102 in response to the inquiry information to the effect that recording to the log is not permitted, the masking processing request is received from the information processing terminal 102. can be done.

As a result, in a conversation with a chatbot, for example, just by performing a simple operation of pressing the secret button 720 (see, for example, FIG. 7) before entering information, the entered information is not left in the log. It becomes possible to

As described above, according to the information processing apparatus 101, the user himself/herself can control the recording of the information input in the chatbot conversation in the log. As a result, it is possible to eliminate anxiety when using a service that involves inputting information that the user does not want to know. In addition, even if the input information is masked, the information on which step of the talk script (scenario) was executed remains in the log, so it can be used for flow line analysis on the operator side.

(Embodiment 2)
Next, the information processing apparatus 101 according to the second embodiment will be described. In the second embodiment, the information processing apparatus 101 receives input information received at a location where information is input in a chatbot conversation, and receives a masking processing request for the input information. Record the number of times Then, a case will be described where the information processing apparatus 101 determines whether or not the input information received at the location is masking processing target information based on the number of times of recording. The same reference numerals are assigned to the same portions as those described in the first embodiment, and illustration and description thereof are omitted.

(Example of functional configuration of information processing apparatus 101)
First, with reference to FIG. 13, a functional configuration example of the information processing apparatus 101 according to the second embodiment will be described.

FIG. 13 is a block diagram of a functional configuration example of the information processing apparatus 101 according to the second embodiment. In FIG. 13, the information processing apparatus 101 includes a reception unit 801, a conversation control unit 802, a recording control unit 803, and a counting unit 1301. Specifically, for example, the reception unit 801 to the recording control unit 803 and the counting unit 1301 cause the CPU 201 to execute programs stored in storage devices such as the memory 202, the disk 204, and the portable recording medium 207 shown in FIG. or by the communication I/F 205, the function is realized. The processing results of each functional unit are stored in a storage device such as the memory 202 or disk 204, for example.

Hereinafter, of the functional units of the information processing apparatus 101 according to the second embodiment, portions different from the functional units of the information processing apparatus 101 according to the first embodiment will be described.

The counting unit 1301 receives the input information accepted at the first part in association with the first part in the chatbot conversation, and records the number of times the masking processing request for the input information is received. Here, the first part is the part where information is input in the chatbot conversation.

For example, in the first part, out of a plurality of steps in the script, a talk process is executed in which input of information and selection of options are accepted, and the input information and information indicating the selected options are spoken. step. The number of times corresponding to the first location is stored, for example, in a number counter table 1400 as shown in FIG.

Here, the contents stored in the number counter table 1400 will be described. The number counter table 1400 is implemented by a storage device such as the memory 202 and disk 204 shown in FIG. 2, for example.

FIG. 14 is an explanatory diagram showing an example of the contents stored in the number counter table 1400. As shown in FIG. In FIG. 14, the number counter table 1400 has fields of script, label, and number of times. By setting information in each field, number information (for example, number information 1400-1 to 1400-4) is stored as a record. do.

Here, script indicates a script (scenario) applied to the chatbot conversation. label indicates the name of the step in which talk processing is performed among the steps in the script. The number of times indicates the number of times the input information accepted in the step is received and the masking processing request for the input information is received.

For example, the number of times information 1400-1 indicates the number of times "5" that the input information received at the step labeled "VRB-1" in the script P100 and the masking processing request for the input information were received.

As an example, in a chatbot conversation based on the talk script P100, it is assumed that the input information accepted in the step labeled "VRB-1" is received and a masking processing request for the input information is received. In this case, counting unit 1301 increments the number of times information 1400-1 corresponding to script “P100” and label “VRB-1” in number counter table 1400. FIG.

When the recording control unit 803 receives the input information received at the first part in the chatbot conversation from the information processing terminal 102, the recording control unit 803, based on the number of times recorded in association with the first part, It is determined whether or not the input information is masking processing target information. Here, the information to be masked is information that is not recorded in the log or that is to be made anonymous when recording a log related to chatbot conversation.

Specifically, for example, the recording control unit 803 determines that the received input information is masking target information when the number of times recorded in association with the first location is greater than or equal to the first threshold value. to decide. The first threshold can be arbitrarily set, and is set, for example, according to the length of the period in which the number of times corresponding to the first location is recorded.

As an example, assume that the number of times corresponding to the first location is recorded for a period of one day. In this case, the first threshold is set to a value of about ten, for example. It is also assumed that the number of times corresponding to the first location is recorded for a period of one week. In this case, the first threshold is set to a value of about 100, for example.

In addition, when the ratio of the number of times of recording to the total number of times of receiving the input information accepted at the first location is equal to or greater than the second threshold value, the recording control unit 803 determines that the received input information is subject to masking processing. You may decide to judge that it is information. The second threshold can be set arbitrarily, and is set to a value of about 0.6 (60%) to 0.8 (80%), for example.

Here, if it is determined that the information is subject to masking processing, the recording control unit 803 performs masking processing on the received input information when recording a log relating to the chatbot conversation. On the other hand, if the recording control unit 803 determines that the information is not masking processing target information, the recording control unit 803 does not perform masking processing on the received input information.

In addition, the conversation control unit 802, when the number of times recorded in association with the first part in the chatbot conversation is greater than or equal to the first threshold value, the information input at the first part Inquiry information as to whether or not recording to the log of is permitted may be output to the information processing terminal 102 .

Then, when the recording control unit 803 receives the input information accepted at the first location from the information processing terminal 102 and receives a masking processing request for the input information, when recording a log related to the chatbot conversation, , the masking process may be performed on the input information.

(Conversation control processing procedure of information processing device 101)
Next, a conversation control processing procedure of the information processing apparatus 101 according to the second embodiment will be explained. In the information processing apparatus 101 according to the second embodiment, for example, after performing the first conversation control process for a predetermined period, the second conversation control process is performed.

First, the first conversation control processing procedure of the information processing apparatus 101 according to the second embodiment will be described with reference to FIGS. 15 and 16. FIG.

15 and 16 are flowcharts showing an example of the first conversation control processing procedure of the information processing apparatus 101 according to the second embodiment. In the flowchart of FIG. 15, first, the information processing apparatus 101 determines whether or not a chat start instruction has been received from the information processing terminal 102 (step S1501).

Here, the information processing apparatus 101 waits to receive a chat start instruction (step S1501: No). When the information processing apparatus 101 receives the chat start instruction (step S1501: Yes), the information processing apparatus 101 refers to the script table 130 and determines the script corresponding to the service ID included in the received chat start instruction as the talk script. (step S1502).

Then, the information processing apparatus 101 uses the determined talk script to start a chatbot conversation on the information processing terminal 102 (step S1503). At this time, the information processing apparatus 101 outputs to the information processing terminal 102 inquiry information as to whether or not to permit recording of the input information in the log. The inquiry information is, for example, the confidential button 720 shown in FIG.

The information processing apparatus 101 then determines whether or not an off-record setting instruction has been received from the information processing terminal 102 (step S1504). If the off-record setting instruction is received from the information processing terminal 102 (step S1504: Yes), the information processing apparatus 101 turns on the off-record mode (step S1505), and proceeds to step S1601 shown in FIG.

On the other hand, if the off-the-record setting instruction has not been received from the information processing terminal 102 (step S1504: No), the information processing apparatus 101 determines whether or not the off-the-record cancellation instruction has been received from the information processing terminal 102 (step S1506). .

Here, if the off-the-record cancellation instruction has not been received from the information processing terminal 102 (step S1506: No), the information processing apparatus 101 proceeds to step S1601 shown in FIG. On the other hand, when the off-the-record cancellation instruction is received from the information processing terminal 102 (step S1506: Yes), the information processing apparatus 101 turns off the off-the-record mode (step S1507), and proceeds to step S1601 shown in FIG.

In the flowchart of FIG. 16, first, the information processing apparatus 101 determines whether or not input information accepted in the chatbot conversation is received from the information processing terminal 102 (step S1601). If the input information has not been received (step S1601: No), the information processing apparatus 101 proceeds to step S1607.

On the other hand, if input information has been received (step S1601: Yes), the information processing apparatus 101 executes response processing based on the received input information (step S1602). The information processing apparatus 101 determines whether the off-record mode is ON (step S1603).

If the off-record mode is ON (step S1603: YES), the information processing apparatus 101 performs masking processing on the received input information and outputs a chat log in which the input information is masked (step S1604). The information processing apparatus 101 then increments the number of times of the corresponding label in the number counter table 1400 (step S1605), and proceeds to step S1607.

On the other hand, if the off-record mode is OFF (step S1603: No), the information processing apparatus 101 outputs a chat log including the received input information (step S1606). The information processing apparatus 101 then determines whether or not to end the chatbot conversation on the information processing terminal 102 (step S1607).

Here, if the conversation is not to end (step S1607: No), the information processing apparatus 101 returns to step S1504 shown in FIG. On the other hand, when ending the conversation (step S1607: Yes), the information processing apparatus 101 ends the series of processes according to this flowchart.

As a result, in response to an off-the-record setting instruction from the information processing terminal 102, it is possible to mask the input information received in the chatbot conversation and output it as a log. In addition, it is possible to count the number of times masking processing requests have been made for the input information received in the steps (places) in the talk script.

Next, a second conversation control processing procedure of the information processing apparatus 101 according to the second embodiment will be described with reference to FIGS. 17 and 18. FIG.

17 and 18 are flowcharts showing an example of the second conversation control processing procedure of the information processing apparatus 101 according to the second embodiment. In the flowchart of FIG. 17, first, the information processing apparatus 101 determines whether or not a chat start instruction has been received from the information processing terminal 102 (step S1701).

Here, the information processing apparatus 101 waits to receive a chat start instruction (step S1701: No). When the information processing apparatus 101 receives the chat start instruction (step S1701: Yes), the information processing apparatus 101 refers to the script table 130 and determines the script corresponding to the service ID included in the received chat start instruction as the talk script. (step S1702).

Then, the information processing apparatus 101 uses the determined talk script to start a chatbot conversation on the information processing terminal 102 (step S1703). Next, the information processing apparatus 101 refers to the number counter table 1400 to obtain the number of times of the label at the current position in the determined talk script (step S1704).

Then, the information processing apparatus 101 determines whether or not the acquired number of times is equal to or greater than the threshold value (step S1705). Here, if the number of times is less than the threshold (step S1705: No), the information processing apparatus 101 turns off the off-record mode (step S1707), and proceeds to step S1801 shown in FIG.

Note that in step S1704, if the number of times of the current position cannot be acquired, the information processing apparatus 101 proceeds to step S1707.

On the other hand, if the number of times is equal to or greater than the threshold (step S1705: Yes), the information processing apparatus 101 turns on the off-record mode (step S1706), and proceeds to step S1801 shown in FIG.

At this time, the information processing apparatus 101 may activate the secret button 720 in the chat screen 700 as shown in FIG. 10, for example. Furthermore, the information processing apparatus 101 may display a message such as "You are in off-the-record mode. The entered information is not recorded on the server." . Also, in the second conversation control process, for example, the secret button 720 may not be displayed in the chat screen 700 .

In the flowchart of FIG. 18, first, the information processing apparatus 101 determines whether or not input information accepted in the chatbot conversation is received from the information processing terminal 102 (step S1801). If the input information has not been received (step S1801: No), the information processing apparatus 101 proceeds to step S1806.

On the other hand, when input information is received (step S1801: Yes), the information processing apparatus 101 executes response processing based on the received input information (step S1802). The information processing apparatus 101 determines whether the off-record mode is ON (step S1803).

Here, if the off-record mode is ON (step S1803: Yes), the information processing apparatus 101 performs masking processing on the received input information and outputs a chat log in which the input information is masked (step S1804). The process moves to step S1806.

On the other hand, if the off-record mode is OFF (step S1803: No), the information processing apparatus 101 outputs a chat log including the received input information (step S1805). The information processing apparatus 101 then determines whether or not to end the chatbot conversation on the information processing terminal 102 (step S1806).

Here, if the conversation is not to end (step S1806: No), the information processing apparatus 101 returns to step S1704 shown in FIG. On the other hand, when ending the conversation (step S1806: YES), the information processing apparatus 101 ends the series of processes according to this flowchart.

As a result, it is possible to perform masking processing and log output of input information received at a step in which the number of past masking processing requests is greater than or equal to a threshold among the steps (places) in the talk script.

As described above, according to the information processing apparatus 101 according to the second embodiment, the input information received at the first place in the chatbot conversation is received in association with the first place in the chatbot conversation. , the number of times the masking process request for the input information is received can be recorded.

As a result, operations for not leaving input information in the log, for example, the number of times the secret button 720 is pressed, can be recorded at locations (steps in the script) where information is input in a chatbot conversation. .

Further, according to the information processing apparatus 101, when the input information received at the first part in the chatbot conversation is received from the information processing terminal 102, the received input information is subjected to masking processing based on the number of recordings. It can be determined whether or not the information is the target information. Then, according to the information processing apparatus 101, when it is determined that the received input information is information to be masked, the information processing apparatus 101 can perform masking processing of not recording the received input information in a log or making it anonymous.

As a result, it is possible to determine whether the information entered at that location is subject to masking processing, considering how many operations have been performed in the past in order not to be recorded in the log at the location where information is input. can do. Therefore, it is possible to prevent the entered information from being recorded in the log without the user having to perform an operation such as pressing the secret button 720 for a portion where a large number of users have performed an operation not to be recorded in the log. Become.

Further, according to the information processing apparatus 101, the information processing apparatus 101 can determine that the received input information is masking processing target information when the number of times of recording is equal to or greater than the threshold value. Further, according to the information processing apparatus 101, when the ratio of the number of times of recording to the total number of times of receiving the input information input to the first portion is equal to or greater than the threshold value, the received input information is masked. can be determined to be information.

This automatically sets the off-the-record mode, which is where the information to be masked is input, taking into account the number and percentage of operations that other users performed to prevent them from being recorded in the log during the aggregation period. You can decide where to do it.

(Embodiment 3)
Next, the information processing apparatus 101 according to the third embodiment will be explained. In the third embodiment, a case will be described where attention is called when personal identification information is included in input information received in a chatbot conversation. The same reference numerals are assigned to the same parts as those described in the first and second embodiments, and illustration and explanation are omitted.

(Example of functional configuration of information processing apparatus 101)
First, with reference to FIG. 19, a functional configuration example of the information processing apparatus 101 according to the third embodiment will be described.

FIG. 19 is a block diagram of a functional configuration example of the information processing apparatus 101 according to the third embodiment. In FIG. 19 , the information processing apparatus 101 includes a reception unit 801 , a conversation control unit 802 , a recording control unit 803 and a determination unit 1901 . Specifically, for example, the reception unit 801 to the recording control unit 803 and the determination unit 1901 cause the CPU 201 to execute programs stored in storage devices such as the memory 202, the disk 204, and the portable recording medium 207 shown in FIG. or by the communication I/F 205, the function is realized. The processing results of each functional unit are stored in a storage device such as the memory 202 or disk 204, for example.

Hereinafter, of the functional units of the information processing apparatus 101 according to the third embodiment, portions different from the functional units of the information processing apparatus 101 according to the first embodiment will be described.

When the input information accepted in the chatbot conversation is received from the information processing terminal 102, the determination unit 1901 determines whether or not the input information includes personal identification information. Here, the personal identification information is information that can identify an individual, such as last name, first name, and e-mail address.

Determination as to whether personal identification information is included in the input information is performed using, for example, a personal identification information master 2000 as shown in FIG. Here, the contents stored in the personal identification information master 2000 will be explained. The personal identification information master 2000 is implemented by a storage device such as the memory 202 and disk 204 shown in FIG. 2, for example.

FIG. 20 is an explanatory diagram showing an example of the contents stored in the personal identification information master 2000. As shown in FIG. In FIG. 20, an individual identification information master 2000 is dictionary information used for determining whether information that can identify an individual is included. Personal identification information master 2000 stores, for example, character strings such as surnames, first names, and domain names.

Specifically, for example, the determination unit 1901 analyzes the input information and determines that the input information includes personal identification information when the character string stored in the personal identification information master 2000 is included. For example, when the input information includes the character string "Sato", it is determined that the input information includes personal identification information.

When the determination unit 1901 determines that the input information includes personal identification information, the conversation control unit 802 outputs to the information processing terminal 102 recommendation information for masking information input in the chatbot conversation. Here, the recommendation information is a message, an icon, or the like for recommending execution of the masking process.

Specifically, for example, the conversation control unit 802 sends recommendation information for masking processing to the information processing terminal in association with inquiry information as to whether or not recording of input information in a log is permitted in a chatbot conversation. 102 may be output. The masking processing request is transmitted from the information processing terminal 102, for example, when an operation is performed on the information processing terminal 102 to the effect that recording to the log is not permitted in response to the inquiry information.

Accordingly, after the input information including the personal identification information is input, the user can select whether or not to perform the masking process on the information input thereafter. A specific example of masking process recommendation information output in association with inquiry information will be described later with reference to FIG. 23 .

Further, the determination unit 1901 may use the machine learning model M to determine whether input information received in a chatbot conversation on the information processing terminal 102 includes a specific word. Here, specific words are, for example, passwords, credit card numbers, and the like. The machine learning model M is a machine learning model for classifying whether an input character string is a specific word.

Specifically, for example, the determination unit 1901 morphologically analyzes the received input information (text data) and divides it into words. Next, the determination unit 1901 uses the machine learning model M to determine whether each divided word is a specific word. When determining that at least one of the words includes the specific word, the determination unit 1901 determines that the input information includes the specific word.

The machine learning model M is generated by machine learning such as deep learning using, for example, character string data labeled to indicate whether or not it is a specific word as learning data (teacher data). More specifically, for example, the machine learning model M can be generated by deep learning using algorithms such as CNN and LSTM.

Note that CNN is an abbreviation for Convolutional Neural Network. LSTM is an abbreviation for Long Short-Term Memory. An example of generating the machine learning model M will be described later with reference to FIG. An example of determination of specific words will be described later with reference to FIG. 22 .

Conversation control section 802 may output alerting information to information processing terminal 102 when determination section 1901 determines that a specific word is included in the input information. Here, the attention calling information is a message, an icon, or the like for calling attention to whether or not the user has entered important information. Specifically, for example, the conversation control unit 802 outputs alert information to the information processing terminal 102 in association with input information determined to include a specific word in a chatbot conversation. may

Also, the conversation control unit 802 may receive new input information from the information processing terminal 102 for input information determined to include a specific word. In this case, conversation control section 802 executes response processing based on new input information received from information processing terminal 102 . Also, the conversation control unit 802 discards (deletes) input information determined to include a specific word, for example. That is, no response processing is executed and no log is recorded for input information determined to include a specific word.

Also, the alert information may be information that recommends execution of masking processing for input information determined to include a specific word. In this case, when the recording control unit 803 receives a masking processing request for input information determined to include a specific word, when recording a log related to a chatbot conversation, the recording control unit 803, when recording the log related to the conversation, A masking process is performed on the input information in the information log.

A specific example of the alerting information that is output when it is determined that the input information includes a specific word will be described later with reference to FIG. 24 .

(Generation example of machine learning model M)
Next, an example of generating the machine learning model M will be described with reference to FIG. 21 . Here, "password" will be described as an example of the specific word.

FIG. 21 is an explanatory diagram showing a specific example of learning data. In FIG. 21, learning data 2100 is a set of labeled teacher data (for example, teacher data 2100-1 to 2100-4). For example, the learning data 2100 includes teacher data regarding each character string of a character string group (reference numeral 2110 in FIG. 21) including a random character string group and an English word group.

The random character string group is a character string group assuming a password, for example, a character string group that uniformly includes arbitrary character strings of patterns 1 to 9 below. Example) is an example of a character string corresponding to each pattern.

Pattern 1: With numbers/with symbols/with capital letters, string length: 4 to 20
Example) w7mG$ijG

Pattern 2: With numbers/No symbols/No capital letters, String length: 4 to 20
Example) a0r37m6b

Pattern 3: No numbers/symbols/no capital letters, string length: 4 to 20
Example) a. opi|/$

Pattern 4: no numbers/no symbols/capital letters, string length: 4 to 20
Example) GqADkYwH

Pattern 5: Numbers/symbols/no capital letters, string length: 4 to 20
Example) |m39)uk#

Pattern 6: with numbers/without symbols/with capital letters, string length: 4 to 20
Example) Dm3fGdQo

Pattern 7: Numbers/symbols/no capital letters, string length: 4 to 20
Example) -z02; sm6

Pattern 8: No numbers/with symbols/with capital letters, string length: 4 to 20
Example) fz. Cs {! ~

Pattern 9: no numbers/no symbols/no capital letters, string length: 4 to 20
Example) miqmxyfi

An English word group is a character string group containing arbitrary English words. The random character string group and the English word group each include about several thousand (eg, 7000 to 8000) character strings.

Each character string (word) is represented by a vector representation (reference numeral 2120 in FIG. 21). For example, each character string is continuous data (character string data) in which vector data representing each character included in each character string are arranged in order of appearance. Each character is represented by, for example, a 128-dimensional one-hot vector in which the column corresponding to the ASCII code of each character is "1" and the other columns are "0". ASCII stands for American Standard Code for Information Interchange.

The label is information indicating whether the character string is a password (reference numeral 2130 in FIG. 21). Here, if the character string is a password, the label "0.0" is given. Also, if the character string is non-password, the label "1.0" is given.

For example, based on the learning data 2100, the determination unit 1901 regards the character string as time-series continuous data, forms RNN/LSTM, and generates a machine learning model M by performing machine learning. The machine learning model M, for example, receives as input continuous data (array) of ASCII characters (vector representation) and outputs a value indicating the degree of password recognition.

The value indicating the degree of password is represented by a numerical value (decimal number) from 0 to 1. Here, the value indicating the degree of being a password ranges from 0 to 1, and the higher the degree of being a password, the lower the value, and the lower the degree of being a password, the higher the value.

Here, an example of password determination will be described with reference to FIG. 22 . Here, it is assumed that the character string to be determined is "hVLVcaWf".

FIG. 22 is an explanatory diagram of an example of password determination. In FIG. 22, continuous data 2200 is a vector representation of the character string "hVLVcaWf". Specifically, the continuous data 2200 is obtained by arranging vector data representing each character included in the character string "hVLVcaWf" in order of appearance.

For example, the character "h" included in the character string "hVLVcaWf" is represented by vector data in which the 104th column corresponding to the ASCII code is "1" and the other columns are "0". Also, the character "V" included in the character string "hVLVcaWf" is represented by vector data in which the 86th column corresponding to the ASCII code is "1" and the other columns are "0".

Here, it is assumed that as a result of inputting continuous data 2200 to the machine learning model M, an output value of "0.00001" is output. In this case, for example, when the output value “0.00001” of the machine learning model M is less than the threshold α, the determination unit 1901 determines that it is a password. The threshold α can be arbitrarily set, and is set to a value of about 0.5, for example. For example, if the threshold value α is “α=0.5”, the determination unit 1901 determines that the character string “hVLVcaWf” is a password because the output value “0.00001” is less than the threshold value α. .

(Specific example of recommendation information for masking processing)
Next, a specific example of masking process recommendation information output in association with inquiry information will be described with reference to FIG. 23 .

FIG. 23 is an explanatory diagram (part 3) showing an example of a chat screen. In FIG. 23 , a chat screen 2300 is an example of an operation screen that displays a chatbot conversation on the information processing terminal 102 . Here, as in the case of the chat screen 700 shown in FIG. 9, a case where some kind of consultation is performed using two-way communication in cooperation with a chatbot will be described as an example.

On the chat screen 2300, for example, a message can be input by selecting a box 2310 through user operation input using the input device 304 shown in FIG. Also, by selecting a send button 2311 on the chat screen 2300, the message input in the box 2310 can be sent to the chatbot side (information processing apparatus 101).

Here, on the chat screen 2300, messages m11 and m13 representing the utterances of the avatar av and a message m12 representing the user's utterances are displayed. The message m11 is an example of a message prompting the user to enter his/her name. Message m12 is an example of the name entered for message m1 of avatar av. The message m13 is an example of a message that prompts the user to input consultation content.

Here, the user's name "Taro Sato" is input prior to inputting the content of the user's consultation. In this case, "Taro Sato" received in the chatbot conversation is transmitted from the information processing terminal 102 to the information processing apparatus 101 as input information. The determination unit 1901 refers to the personal identification information master 2000 and determines whether personal identification information is included in the input information “Taro Sato”.

Here, since the input information "Sato Taro" includes "Sato" stored in the personal identification information master 2000, it is determined that the input information "Sato Taro" includes personal identification information. In this case, the conversation control unit 802 outputs masking processing recommendation information to the information processing terminal 102 in association with inquiry information as to whether or not recording of the input information in the log is permitted.

That is, when information that can identify an individual is input in a chatbot conversation, the information that is input thereafter is recorded on the server (for example, the information processing device 101) in a state that allows the individual to be identified. will be Therefore, the information processing apparatus 101 recommends masking processing for information that is subsequently input in the chatbot conversation.

In the example of FIG. 23, a recommendation message 2330 is displayed in association with a secret button 2320 on the chat screen 2300 . A secret button 2320 is an example of inquiry information as to whether or not to permit recording of input information in the log. A recommendation message 2330 is an example of masking process recommendation information.

According to the recommendation message 2330, if the user presses the secret button 2320, the user can understand that the input information will not be recorded in the log. For example, it is assumed that the user selects the secret button 2320 on the chat screen 2300 because the user does not want to leave his or her consultation contents in the log. In this case, an off-record setting instruction is transmitted from the information processing terminal 102 to the information processing apparatus 101, and the off-record mode is set (ON).

(Specific example of alert information)
Next, with reference to FIG. 24, a specific example of attention calling information that is output when it is determined that the input information contains a specific word will be described.

FIG. 24 is an explanatory diagram (part 4) showing an example of a chat screen. In FIG. 24 , a chat screen 2400 is an example of an operation screen that displays a chatbot conversation on the information processing terminal 102 . Here, as in the case of the chat screen 700 shown in FIG. 9, a case where some kind of consultation is performed using two-way communication in cooperation with a chatbot will be described as an example.

Here, on the chat screen 2400, messages m21 and m23 representing the utterances of the avatar av and a message m22 representing the user's utterances are displayed. Here, it is assumed that the user's consultation content is input in box 2410 in response to message m23 of avatar av, and send button 2411 is selected.

In this case, the content of consultation input in box 2410 is transmitted from information processing terminal 102 to information processing apparatus 101 as input information. The determination unit 1901 divides the input information (consultation content) into words, and uses the machine learning model M to determine whether each divided word includes a password.

For example, taking the part of the consultation content entered in the box 2410, "By the way, your password is hVLVcaWf." ”. Then, the word "hVLVcaWf" is determined to be the password.

In this case, conversation control unit 802 outputs alert information to information processing terminal 102 in association with input information determined to include a password. For example, in a conversation with a chatbot, if a sentence containing a password is entered as a consultation content of the user, it is possible that the password was entered by mistake.

In the example of FIG. 24 , alert information 2430 is displayed in association with a secret button 2420 provided next to a box 2410 on a chat screen 2400 . The secret button 2420 is an example of inquiry information as to whether or not to permit recording of input information in the log. The alert information 2430 is an example of information for alerting whether important information has been input and recommending execution of masking processing for the input information.

According to the alert information 2430, the user can check whether important information has been entered. For example, when the user realizes that he or she has entered a password by mistake, he or she can reenter the content of the consultation in the box 2410 and select the send button 2411 to send the new input information from the information processing terminal 102. It can be transmitted to the information processing device 101 .

In this case, conversation control section 802 executes response processing based on new input information received from information processing terminal 102 . In addition, conversation control section 802 discards input information determined to include a password. In addition, the recording control unit 803 does not log input information determined to include a password.

If the user determines that there is no problem even if the password is included, the input information can be confirmed by selecting the send button 2411 again without re-entering the consultation content. In this case, response processing based on the input information including the password is executed.

By selecting the secret button 2420 and then selecting the send button 2411 again, the user can confirm the input information and set (ON) the off-record mode. In this case, input information is masked when recording a log of a chatbot conversation.

(Conversation control processing procedure of information processing device 101)
Next, a conversation control processing procedure of the information processing apparatus 101 according to the third embodiment will be explained. First, the first conversation control processing procedure of the information processing apparatus 101 according to the third embodiment will be described with reference to FIGS. 25 and 26. FIG. In the first conversation control process, masking process is recommended when personal identification information is included in the input information accepted in the chatbot conversation.

25 and 26 are flowcharts showing an example of the first conversation control processing procedure of the information processing apparatus 101 according to the third embodiment. In the flowchart of FIG. 25, first, the information processing apparatus 101 determines whether or not a chat start instruction has been received from the information processing terminal 102 (step S2501).

Here, the information processing apparatus 101 waits to receive a chat start instruction (step S2501: No). When the information processing apparatus 101 receives the chat start instruction (step S2501: Yes), the information processing apparatus 101 refers to the script table 130 and determines the script corresponding to the service ID included in the received chat start instruction as the talk script. (step S2502).

Then, the information processing apparatus 101 uses the determined talk script to start a chatbot conversation on the information processing terminal 102 (step S2503). At this time, the information processing apparatus 101 outputs to the information processing terminal 102 inquiry information as to whether or not to permit recording of the input information in the log. The inquiry information is, for example, the secret button 2320 shown in FIG.

Next, the information processing apparatus 101 determines whether personal identification information has been input in the chatbot conversation (step S2504). Here, if personal identification information has not been input (step S2504: No), the information processing apparatus 101 proceeds to step S2506.

On the other hand, if personal identification information has already been input (step S2504: YES), the information processing apparatus 101 associates the input information with the inquiry information as to whether or not to permit recording of the input information in the log in the chatbot conversation. , the masking processing recommendation information is output to the information processing terminal 102 (step S2505).

The information processing apparatus 101 determines whether or not an off-record setting instruction has been received from the information processing terminal 102 (step S2506). If the off-record setting instruction has not been received from the information processing terminal 102 (step S2506: No), the information processing apparatus 101 proceeds to step S2601 shown in FIG.

On the other hand, when the off-record setting instruction is received from the information processing terminal 102 (step S2506: Yes), the information processing apparatus 101 turns on the off-record mode (step S2507), and proceeds to step S2601 shown in FIG.

In the flowchart of FIG. 26, first, the information processing apparatus 101 determines whether or not input information accepted in the chatbot conversation is received from the information processing terminal 102 (step S2601). Here, if the input information has not been received (step S2601: No), the information processing apparatus 101 proceeds to step S2609.

On the other hand, if input information has been received (step S2601: Yes), the information processing apparatus 101 executes response processing based on the received input information (step S2602). The information processing apparatus 101 determines whether the off-record mode is ON (step S2603).

If the off-record mode is ON (step S2603: YES), the information processing apparatus 101 performs masking processing on the received input information and outputs a chat log in which the input information is masked (step S2604). The information processing apparatus 101 then turns off the off-record mode (step S2605), and proceeds to step S2607.

On the other hand, if the off-record mode is OFF (step S2603: No), the information processing apparatus 101 outputs a chat log including the received input information (step S2606). The information processing apparatus 101 then determines whether or not the received input information includes personal identification information (step S2607).

If the input information does not include personal identification information (step S2607: No), the information processing apparatus 101 proceeds to step S2609. On the other hand, if personal identification information is included in the input information (step S2607: Yes), the information processing apparatus 101 records that personal identification information has been input (step S2608).

The information processing apparatus 101 then determines whether or not to end the chatbot conversation on the information processing terminal 102 (step S2609). Here, if the conversation is not to be ended (step S2609: No), the information processing apparatus 101 returns to step S2504 shown in FIG. On the other hand, when ending the conversation (step S2609: Yes), the information processing apparatus 101 ends the series of processes according to this flowchart.

As a result, when personal identification information is included in the input information received in the chatbot conversation on the information processing terminal 102, it is possible to recommend masking the information that will be subsequently input in the chatbot conversation.

Next, a second conversation control processing procedure of the information processing apparatus 101 according to the third embodiment will be described with reference to FIG. In the second conversation control process, if the input information accepted in the chatbot conversation includes a password, the attention calling information is output. Also, in FIG. 27, the description of the process of recording a log regarding the chatbot conversation is omitted.

FIG. 27 is a flowchart showing an example of the second conversation control processing procedure of the information processing apparatus 101 according to the third embodiment. In the flowchart of FIG. 27, first, the information processing apparatus 101 determines whether or not a chat start instruction has been received from the information processing terminal 102 (step S2701).

Here, the information processing apparatus 101 waits to receive a chat start instruction (step S2701: No). When the information processing apparatus 101 receives the chat start instruction (step S2701: Yes), the information processing apparatus 101 refers to the script table 130 and determines the script corresponding to the service ID included in the received chat start instruction as the talk script. (step S2702).

Then, the information processing apparatus 101 uses the determined talk script to start a chatbot conversation on the information processing terminal 102 (step S2703). At this time, the information processing apparatus 101 may output to the information processing terminal 102 inquiry information as to whether or not to permit recording of the input information in the log. The inquiry information is, for example, the secret button 2420 shown in FIG.

Next, the information processing apparatus 101 determines whether or not the input information accepted in the chatbot conversation is received from the information processing terminal 102 (step S2704). If the input information has not been received (step S2704: No), the information processing apparatus 101 proceeds to step S2711.

On the other hand, when input information is received (step S2704: Yes), the information processing apparatus 101 executes password detection processing for the received input information (step S2705). The password detection process is a process of detecting a password included in input information using a machine learning model M. FIG.

The information processing apparatus 101 then determines whether a password has been detected (step S2706). Here, if the password is not detected (step S2706: No), the information processing apparatus 101 executes response processing based on the received input information (step S2707), and proceeds to step S2711.

On the other hand, if a password is detected (step S2706: YES), the information processing apparatus 101 outputs alert information to the information processing terminal 102 (step S2708). The information processing apparatus 101 then determines whether new input information has been received from the information processing terminal 102 for the input information for which the password has been detected (step S2709).

If new input information has not been received (step S2709: No), the information processing apparatus 101 proceeds to step S2707. On the other hand, if new input information has been received (step S2709: YES), the information processing apparatus 101 executes response processing based on the received new input information (step S2710).

The information processing apparatus 101 then determines whether or not to end the chatbot conversation on the information processing terminal 102 (step S2711). Here, if the conversation is not to end (step S2711: No), the information processing apparatus 101 returns to step S2704. On the other hand, when ending the conversation (step S2711: Yes), the information processing apparatus 101 ends the series of processes according to this flowchart.

Accordingly, when a password is included in input information received in a chatbot conversation, it is possible to call attention to whether or not the password is entered by mistake.

As described above, according to the information processing apparatus 101 according to the third embodiment, when input information accepted in a chatbot conversation is received from the information processing terminal 102, whether or not the input information includes personal identification information. It is possible to determine whether or not Then, according to the information processing device 101 , when it is determined that the personal identification information is included, the information that recommends masking processing for the information input in the chatbot conversation can be output to the information processing terminal 102 .

As a result, when personal identification information is included in the input information received in the chatbot conversation on the information processing terminal 102, it is possible to recommend masking the information that will be subsequently input in the chatbot conversation. For example, once personally identifiable information is entered, any information subsequently entered will be logged in an identifiable manner. For this reason, when information that can identify an individual is input, masking processing is recommended for information that is input after that.

Further, according to the information processing device 101, using a machine learning model (M) for classifying whether an input character string is a specific word, in a chatbot conversation on the information processing terminal 102, It can be determined whether or not the received input information includes a specific word. Then, according to the information processing apparatus 101 , when it is determined that the input information includes a specific word, the information processing apparatus 101 can output alert information to the information processing terminal 102 .

As a result, when input information received in a chatbot conversation includes a specific word such as a password, it is possible to call attention to whether or not unnecessary information has been input by mistake.

For these reasons, according to the information processing apparatus 101 according to the first to third embodiments, it is possible to smoothly use services using chatbots. Embodiments 1 to 3 may be combined without contradiction. For example, the information processing apparatus 101 according to the first embodiment may have the functions of the information processing apparatus 101 according to the second embodiment and the functions of the information processing apparatus 101 according to the third embodiment.

The conversation control method described in this embodiment can be implemented by executing a prepared program on a computer such as a personal computer or a workstation. This conversation control program is recorded in a computer-readable recording medium such as a hard disk, flexible disk, CD-ROM, DVD, USB memory, etc., and is executed by being read from the recording medium by a computer. Also, the conversation control program may be distributed via a network such as the Internet.

Further, the information processing apparatus 101 described in the present embodiment can also be realized by application specific ICs such as standard cells and structured ASICs (Application Specific Integrated Circuits), and PLDs (Programmable Logic Devices) such as FPGAs.

Further, the following additional remarks are disclosed with respect to the above-described embodiment.

(Appendix 1) receiving input information accepted in a chatbot conversation from an information processing terminal, executing response processing based on the input information,
When a request for masking processing of the input information is received, when recording a log regarding the conversation in the chatbot, a masking processing is performed to not record or anonymize the log regarding the input information among the logs regarding the conversation. ,
A conversation control program characterized by causing a computer to execute processing.

(Appendix 2) causing the computer to execute a process of outputting inquiry information as to whether or not to permit recording of input information in a log in the chatbot conversation to the information processing terminal;
The masking processing request is transmitted from the information processing terminal when an operation is performed in the information processing terminal to the effect that recording in the log is not permitted in response to the inquiry information. 1. A conversation control program according to claim 1.

(Appendix 3) In association with the first part in the chatbot conversation, the input information accepted at the first part is received, and the number of times the masking processing request for the input information is received is recorded. ,
When the input information received at the first part in the chatbot conversation is received from the information processing terminal, whether or not the received input information is masking processing target information based on the recorded number of times. determine whether
When it is determined that the information is subject to the masking process, executing a masking process of not recording the received input information in a log or anonymizing it;
The conversation control program according to appendix 1, characterized by causing the computer to execute processing.

(Appendix 4) The process of determining
When the number of recorded times is a threshold value or more, or when the ratio of the number of times to the total number of times the input information input to the first location is received is a threshold value or more, the received input information is 3. The conversation control program according to appendix 3, wherein the information is determined to be the masking target information.

(Appendix 5) When receiving input information received in a chatbot conversation from an information processing terminal, determining whether or not the input information includes personal identification information,
When it is determined that the personal identification information is included, outputting information to the information processing terminal that recommends masking processing for the information input in the conversation with the chatbot;
The conversation control program according to appendix 1, characterized by causing the computer to execute processing.

(Additional remark 6) In the chatbot conversation, the computer outputs the recommended information to the information processing terminal in association with inquiry information as to whether or not to permit recording of input information in a log. and run
The masking processing request is transmitted from the information processing terminal when an operation is performed on the information processing terminal to the effect that recording in the log is not permitted in response to the inquiry information. 5. The conversation control program according to 5.

(Appendix 7) Using a machine learning model for classifying whether the input character string is a specific word, the specific word is included in the input information received in the chatbot conversation on the information processing terminal. determine whether it is included,
When it is determined that the specific word is included in the input information, outputting alert information to the information processing terminal;
The conversation control program according to appendix 1, characterized by causing the computer to execute processing.

(Appendix 8) The machine learning model combines character string data labeled to indicate that it is the specific word and character string data labeled to indicate that it is not the specific word as teacher data. The conversation control program according to appendix 7, which is generated by performing machine learning as

(Additional remark 9) receiving input information accepted in the chatbot conversation from the information processing terminal, executing response processing based on the input information,
When a request for masking processing of the input information is received, when recording a log regarding the conversation in the chatbot, a masking processing is performed to not record or anonymize the log regarding the input information among the logs regarding the conversation. ,
A conversation control method characterized in that processing is executed by a computer.

(Appendix 10) A conversation control unit that, when receiving input information accepted in a chatbot conversation from an information processing terminal, executes a response process based on the input information;
When a request for masking processing of the input information is received, when recording a log regarding the conversation in the chatbot, a masking processing is performed to not record or anonymize the log regarding the input information among the logs regarding the conversation. a recording control unit;
An information processing device comprising:

100 Conversation Control System 101 Information Processing Device 102 Information Processing Terminal 103 Administrator Terminal 110 Network 120 FAQ Master 130 Script Table 140 Chat Log DB
200,300 Bus 201,301 CPU
202, 302 memory 203 disk drive 204 disk 205, 305 communication I/F
206 portable recording medium I/F
207 portable recording medium 303 display 304 input device 306 camera 307 speaker 308 microphone 309 GPS unit 700, 2300, 2400 chat screen 801 reception unit 802 conversation control unit 803 recording control unit 1301 counting unit 1400 frequency counter table 1901 determination unit 2000 individual identification Information master 2100 Learning data

Claims (7)

receiving input information accepted in a chatbot conversation from an information processing terminal, executing response processing based on the input information;
When a request for masking processing of the input information is received, when recording a log regarding the conversation in the chatbot, a masking processing is performed to not record or anonymize the log regarding the input information among the logs regarding the conversation. ,
A conversation control program characterized by causing a computer to execute processing. causing the computer to execute a process of outputting to the information processing terminal inquiry information as to whether or not to permit recording of input information in a log in the chatbot conversation;
The masking processing request is transmitted from the information processing terminal when an operation is performed on the information processing terminal to the effect that recording in the log is not permitted in response to the inquiry information. Item 1. A conversation control program according to item 1. receiving the input information accepted at the first place in association with the first place in the conversation in the chatbot, and recording the number of times the masking processing request for the input information is received;
When the input information received at the first part in the chatbot conversation is received from the information processing terminal, whether or not the received input information is masking processing target information based on the recorded number of times. determine whether
When it is determined that the information is subject to the masking process, the received input information is not recorded in a log or is masked to be anonymized.
2. The conversation control program according to claim 1, causing said computer to execute processing. When input information accepted in a chatbot conversation is received from an information processing terminal, determining whether or not the input information includes personal identification information,
When it is determined that the personal identification information is included, outputting information to the information processing terminal that recommends masking processing for the information input in the conversation with the chatbot;
2. The conversation control program according to claim 1, causing said computer to execute processing. Using a machine learning model for classifying whether an input character string is a specific word or not, whether or not the specific word is included in the input information received in the chatbot conversation on the information processing terminal determine whether
When it is determined that the specific word is included in the input information, outputting alert information to the information processing terminal;
2. The conversation control program according to claim 1, causing said computer to execute processing. receiving input information accepted in a chatbot conversation from an information processing terminal, executing response processing based on the input information;
When a request for masking processing of the input information is received, when recording a log regarding the conversation in the chatbot, a masking processing is performed to not record or anonymize the log regarding the input information among the logs regarding the conversation. ,
A conversation control method characterized in that processing is executed by a computer. a conversation control unit that, when receiving input information accepted in a chatbot conversation from an information processing terminal, executes a response process based on the input information;
When a request for masking processing of the input information is received, when recording a log regarding the conversation in the chatbot, a masking processing is performed to not record or anonymize the log regarding the input information among the logs regarding the conversation. a recording control unit;
An information processing device comprising: