AI system for collecting and sharing dialogue information and structure it

Abstract

AI system for collecting and structuring dialogue information to efficiently share it among members. [Solution] A dialogue information aggregation and structuring sharing AI system comprising: a dialogue unit that provides dialogue threads to terminals used by members belonging to a predetermined group, or that enables dialogue between members and between system members using a large-scale language model via information transmission and information acquisition means installed in the space where the members are present, and acquires dialogue information between the members and between the system members; an analysis unit that analyzes the dialogue information and generates structured knowledge from the dialogue information; and a storage unit that stores the knowledge, wherein the dialogue unit can read the knowledge from the storage unit and convey the content of the knowledge to the members during dialogue between the system members.

Description

Technical Field

[0001] The present invention relates to an interactive information integration and structured sharing AI system that integrates interactive information and shares it among members.

Background Art

[0002] In recent years, attempts have been made to improve the productivity of specific groups such as companies by utilizing generative AI such as large language models. For example, a response generation device that utilizes multiple interactive AIs has been proposed (see Patent Document 1, etc.). In addition, there is also a trend to apply the technology used in so-called social networking services (SNS) that generate chat threads and the like as communication tools within a specific group.

[0003] On the other hand, in activities by specific groups such as companies, insufficient information sharing among the constituent members within the group may have a serious adverse impact on the productivity of the entire group. For example, when Department A in a company conducts an investigation on problem X and obtains conclusion Y, it is extremely inefficient for another Department B in the same company to conduct an investigation on the same problem X without knowing the investigation results of Department A that obtained conclusion Y. Also, even in daily conversations among members where minutes are not created, there may be information exchanges that are valuable for sharing with other members. However, in daily conversations in companies and the like, most are forgotten without being recorded, and in fact, it is rarely converted into information in a form that can be easily shared like minutes.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] This invention relates to an AI system for accumulating and structuring dialogue information, which collects dialogue information between predetermined members belonging to a company or similar organization, and between members and a system, and efficiently shares this information among the members. [Means for solving the problem]

[0006] The AI ​​system for accumulating and structuring dialogue information according to the present invention (hereinafter also simply referred to as the "dialogue information accumulation and sharing system") provides dialogue threads to terminals used by members belonging to a predetermined group, or establishes dialogue between members and between system members using a large-scale language model via information transmission and information acquisition means installed in the space where the members are located, and acquires dialogue information between the members and between the system members, and includes a dialogue unit that acquires dialogue information between the members and between the system members, An analysis unit analyzes the aforementioned dialogue information and generates structured knowledge from the aforementioned dialogue information, It has a storage unit for storing the aforementioned knowledge, The dialogue unit can read the knowledge from the storage unit and transmit the knowledge to the members during a dialogue between the system members.

[0007] Such a dialogue information aggregation and sharing system captures conversations between members and between system members, and stores and stores this information as structured knowledge. This allows for the retrieval of appropriate knowledge in system-member conversations and its transmission to members. Therefore, such a dialogue information aggregation and sharing system can extract and save useful knowledge from everyday conversations that would otherwise be lost without being recorded, provide that knowledge to members in appropriate situations, promote information sharing within the group, and improve the organization's work efficiency.

[0008] Furthermore, for example, when generating the knowledge, the analysis unit may associate the dialogue information with the member who initiated the dialogue information and structure the dialogue information.

[0009] By generating and accumulating knowledge by associating dialogue information with members, it becomes possible to accurately analyze the extent to which that knowledge is shared within the group. Furthermore, by accumulating knowledge that links dialogue information with members, it becomes easier to analyze information about members' skills, such as the richness of the information they possess and the accuracy of their information dissemination. In addition, when the system conveys knowledge to members, it can convey information about the person (member) who possesses the information, either in addition to or instead of conveying the content of the information. Such an analytical unit can more effectively promote information sharing within the group and improve the work efficiency of the organization.

[0010] Furthermore, for example, when generating the knowledge, the analysis unit may extract and classify predetermined information from a plurality of dialogue pieces and structure the dialogue pieces.

[0011] The analysis department extracts and classifies specific information to structure the dialogue data, enabling more effective sharing of the generated knowledge within the organization. For example, the analysis department can extract information such as the purpose (agenda) of the conversation, the starting point (state before the conversation), the ending point (state after the conversation), and the conclusion of the conversation, and then structure the dialogue data.

[0012] Furthermore, for example, the analysis unit determines whether the dialogue between the members and between the system members is a Type 1 information exchange that satisfies predetermined conditions, or a Type 2 information exchange that does not satisfy predetermined conditions. When generating the aforementioned knowledge, the dialogue information may be associated with information indicating whether the dialogue information is the result of a first-type or second-type information exchange, and the dialogue information may be structured accordingly.

[0013] Such an analysis unit can, for example, classify conversations between members and between system members into those where the purpose of the conversation was achieved and those where it was not, and then process each appropriately.

[0014] Furthermore, for example, the analysis unit may determine whether the dialogue between the members and between the system members is a Type 1 information exchange that satisfies predetermined conditions, or a Type 2 information exchange that does not satisfy predetermined conditions. When the analysis unit detects a second type information exchange from the dialogue information, it may use the dialogue unit to propose to the members of the dialogue that an extended dialogue thread be provided, which is a dialogue means for changing the second type information exchange to a first type information exchange. The analysis unit may cause the dialogue unit to generate the extended dialogue thread.

[0015] Such an analysis unit can, for example, extract conversations that have not achieved their objectives and generate extended dialogue threads to achieve those objectives, thereby more effectively promoting information sharing within a group and improving the organization's work efficiency.

[0016] Furthermore, for example, the analysis unit may analyze the dialogue information between members and between system members in the extended dialogue thread, extract progress information regarding the progress of the task of changing the second type of information exchange to the first type of information exchange, structure the progress information in relation to the purpose of generating the extended dialogue thread, and generate the knowledge.

[0017] According to such an analysis department, for example, it is possible to identify conversations whose objectives have not been achieved, manage progress information on tasks that will achieve those objectives, and share progress among the appropriate members.

[0018] Furthermore, for example, when the analysis unit proposes to the members of the dialogue that an extended dialogue thread be established, it also proposes adding additional members other than the members of the dialogue to the extended dialogue thread, The dialogue unit may generate the extended dialogue thread in which the additional member has been added to the dialogue members.

[0019] When generating an extended dialogue thread, such an analysis unit can promote more effective information sharing within the group and improve the work efficiency of the organization by prompting the addition of appropriate collaborators as dialogue members.

[0020] In addition, the dialogue information integration and sharing system according to the present invention further includes a human resource label data generation unit that extracts the skills of the member who issued the dialogue information based on the dialogue information and generates human resource label data in which the member and the skills are associated with each other, and a human resource data storage unit that stores the human resource label data.

[0021] By having such a human resource label data generation unit and a human resource data storage unit, the dialogue information integration and sharing system can integrate information on the skills of the members included in a predetermined group and can provide useful information regarding team formation, for example, to the members or the administrator.

[0022] In addition, for example, the analysis unit may propose to add, as the additional member, a member having the skills useful for changing the second type of information exchange to the first type of information exchange by reading the human resource label data from the human resource data storage unit and adding it to the extended dialogue thread

[0023] Such an analysis unit can recommend a collaborator having appropriate skills as an additional member when generating an extended dialogue thread.

Brief Description of Drawings

[0024] [Figure 1] FIG. 1 is a conceptual diagram showing the functions of a dialogue information integration and sharing system according to an embodiment of the present invention. [Figure 2] FIG. 2 is a functional block diagram of the dialogue information integration and sharing system shown in FIG. 1. [Figure 3] FIG. 3 is a flowchart showing an example of the processing performed in the dialogue information integration and sharing system shown in FIGS. 1 and 2. [Figure 4]Figure 4 is a conceptual diagram illustrating an example of dialogue, knowledge generation, and personnel information generation that take place in the dialogue information accumulation and sharing system shown in Figures 1 and 2. [Figure 5] Figure 5 is a conceptual diagram illustrating another example of dialogue, knowledge generation, and personnel information generation that take place in the dialogue information aggregation and sharing system shown in Figures 1 and 2. [Modes for carrying out the invention]

[0025] Figure 1 is a conceptual diagram showing the relationship and functions of peripheral devices of an interactive information aggregation and sharing system 10 according to one embodiment of the present invention. The interactive information aggregation and sharing system 10 consists of a server device and peripheral devices connected thereto. The server device constituting the interactive information aggregation and sharing system 10 has an internal configuration and functions similar to a general computer, including an arithmetic unit, a storage unit (including the storage unit 40), a display unit, an input unit, and a communication unit. In the description of the interactive information aggregation and sharing system 10, the commonalities with a general computer will be omitted. Furthermore, the server device constituting the interactive information aggregation and sharing system 10 may be cloud-based or on-premise, and is not particularly limited.

[0026] Figure 2 is a functional block diagram of the dialogue information aggregation and sharing system shown in Figure 1. The dialogue information aggregation and sharing system 10, in general terms, includes a dialogue unit 20, an analysis unit 30, a storage unit 40, a personnel label data generation unit 50, a personnel data storage unit 60, and so on.

[0027] As shown in Figures 1 and 2, the dialogue unit 20 of the dialogue information aggregation and sharing system 10 provides dialogue threads 22 to terminals 71-74 used by members belonging to a predetermined group. By providing dialogue threads 22, the dialogue unit 20 enables dialogue between members and between the system and members using a large-scale language model. The dialogue unit 20 also acquires dialogue information between members and between the system and members in the dialogue threads 22.

[0028] Here, the means that the dialogue unit 20 provides for facilitating dialogue between members and between system members using a large-scale language model are not limited to chat threads used in SNS or messaging applications. For example, conference systems and dialogue systems that utilize information transmission and acquisition means installed in the space where members are present, such as communication terminals (including personal computers and smartphones), speakers (including smart speakers), and cameras, can also constitute the dialogue unit 20.

[0029] The dialogue threads 22 of the dialogue information aggregation and sharing system 10 are used as a means of information transmission (communication) between members belonging to a predetermined group, such as a company, organization, or school. As shown in Figure 1, each member belonging to the group can access the dialogue unit 20 via a network communication network (not shown) using member terminals 71-74, which consist of communication terminals such as personal computers and smartphones, and can input text or make voice calls. In this way, the dialogue unit 20 enables dialogue between members belonging to the predetermined group.

[0030] Furthermore, the dialogue unit 20 has a chat response unit 21 that can interact with members using a large-scale language model. The dialogue unit 20 interacts with members accessing the dialogue thread 22 via the chat response unit 21. The chat response unit 21 can respond to response requests (questions, etc.) from members by obtaining the input content of the dialogue thread 22. In addition, the dialogue unit 20, in cooperation with the analysis unit 30 described later, can check (monitor) the dialogue between members in the dialogue thread 22 and interrupt the conversation between members via the chat response unit 21 at a predetermined timing (see steps S003, S004, S007, etc. in Figure 3).

[0031] The extended dialogue thread 24 of the dialogue unit 20, like the dialogue thread 22, is a chat thread that enables dialogue between members and between the system and members using a large-scale language model. However, while the dialogue thread 22 is mainly generated by requests from members (including group administrators) to the system (dialogue unit 20), the extended dialogue thread 24 is mainly generated by a system in which the dialogue unit 20 and the analysis unit 30 work together, making proposals to members. Details of the extended dialogue thread 24 will be described later.

[0032] The analysis unit 30 of the dialogue information accumulation and sharing system 10 analyzes the dialogue information acquired by the dialogue unit 20 via the dialogue thread 22, etc., and generates structured knowledge from the dialogue information. Here, knowledge is not the dialogue information itself (words entered as text or voice by members, etc.) alone, but rather the result of the analysis unit 30 analyzing the dialogue information (words) and surrounding information such as the composition of the thread members, classifying, summarizing, or extracting parts of the dialogue information, processing the dialogue information, or adding information from the analysis results to the dialogue information.

[0033] Specific examples of knowledge will be discussed later.

[0034] The analysis unit 30 includes a knowledge generation unit 32, a thread management unit 34, and an additional member suggestion unit 36. The knowledge generation unit 32 generates knowledge by analyzing the dialogue information of the dialogue thread 22 and the extended dialogue thread 24. The knowledge generation unit 32 mainly structures the dialogue information in relatively short units (small amount of text), such as single statements (text input) or a single round of questions and answers, and generates knowledge.

[0035] The thread management unit 34, like the knowledge generation unit 32, analyzes dialogue information to generate knowledge, and also analyzes the content of the dialogue thread 22, proposing and generating extended dialogue threads 24 to the members of the dialogue thread 22 via the dialogue unit 20. The thread management unit 34 structures dialogue information in relatively long units (large amount of text), such as a series of conversations included in the entire meeting, and generates knowledge such as meeting conclusions, meeting summaries, and task progress management.

[0036] When the analysis unit 30 and the dialogue unit 20 propose to the members of the dialogue (the constituent members of the dialogue thread 22) that an extended dialogue thread 24 be established, the additional member proposal unit 36 ​​proposes adding additional members other than the dialogue members (see Figure 5) to the dialogue members of the extended dialogue thread 24. After obtaining approval from the dialogue members, the dialogue unit 20 generates an extended dialogue thread with the proposed members added to the dialogue members.

[0037] Furthermore, the additional member proposal unit 36 ​​of the analysis unit 30 can read personnel label data (see Figures 4 and 5) from the personnel data storage unit 60, which will be described later, and propose adding members with skills useful for achieving the purpose of the conversation, etc., to the extended dialogue thread as additional members.

[0038] The memory unit 40 stores the knowledge generated by the analysis unit 30. The knowledge stored in the memory unit 40 is read out as needed by the chat response unit 21 of the dialogue unit 20. That is, the dialogue unit 20 can read knowledge from the memory unit 40 during dialogues between system members in a dialogue thread 22, etc., and communicate the content of the read knowledge to the members constituting the dialogue thread 22, etc. This allows the dialogue information aggregation and sharing system 10 to promote information sharing among members of a predetermined group and improve the work efficiency of the organization. The memory unit 40 classifies and stores knowledge for each predetermined group and for each predetermined internal group within that group. This allows the dialogue information aggregation and sharing system 10 to prevent unintended dissemination and leakage of information.

[0039] The personnel label data generation unit 50 extracts information about the skills of the member who initiated the dialogue information based on the dialogue information such as the dialogue thread 22 and the extended dialogue thread 24, and generates personnel label data that associates the member with their skills. Here, the personnel label data can include not only information about the ability to perform a given task (qualifications, skills), but also any information related to the ability to perform the task, such as whether or not the member has knowledge, whether or not they have a network of contacts, and their likes and dislikes.

[0040] Furthermore, the personnel label data generation unit 50 may also extract information about the skills of individuals other than the party who initiated the dialogue (hearsay information about skills) from the dialogue information and generate information about the skills of those individuals.

[0041] The personnel data storage unit 60 stores information such as personnel label data generated by the personnel label data generation unit 50. The information stored in the personnel data storage unit 60 is referenced by the additional member proposal unit 36 ​​when deciding on additional members, and is also referenced by the chat response unit 21 of the dialogue unit 20 when proposing members for a given project in response to requests from members of the dialogue thread 22, etc.

[0042] Furthermore, the personnel label data generated by the personnel label data generation unit 50 can also be considered a type of knowledge generated from dialogue information.

[0043] The following will explain, with specific examples, how the analysis department 30 generates and uses knowledge (such as communicating it to members).

[0044] The knowledge generation unit 32 and thread management unit 34 of the analysis unit 30 shown in Figure 1 may associate dialogue information with the member who initiated the dialogue information and structure the dialogue information when generating knowledge. By generating and accumulating knowledge by associating dialogue information with the member who initiated that information in the dialogue thread 22, the dialogue information accumulation and sharing system 10 can accurately analyze to what extent a given piece of knowledge is shared within a given group. Furthermore, by accumulating knowledge that associates dialogue information with the member who initiated it, it is possible to analyze the richness of information held by a given member and the accuracy of information disseminated by that member. In addition, when the chat response unit 21 conveys the content of the knowledge to a member, it can convey information about the person (member) who possesses the information (e.g., "△△ knows about ○○") instead of conveying the content of the information (e.g., "○○ is ××"). Such an analysis unit 30 can more effectively promote information sharing within a group and improve the work efficiency of the organization.

[0045] Furthermore, the knowledge generation unit 32 and thread management unit 34 of the analysis unit 30 can generate knowledge by extracting and classifying predetermined information from dialogue information and structuring the dialogue information. For example, the knowledge generation unit 32 and thread management unit 34 may extract the agenda (purpose) of the conversation from the content of the statements made immediately after the start of the dialogue thread 22 and generate knowledge (see Figure 3, step S002). Alternatively, the thread management unit 34 of the analysis unit 30 may extract the conclusion of the conversation from a series of dialogue information in the thread, or organize the purpose, start point, and end point of the conversation and generate knowledge (see Figure 3, step S005).

[0046] Furthermore, the dialogue information aggregation and sharing system 10 can perform more advanced processing by linking the communication function of the dialogue unit 20 with the knowledge generation function of the analysis unit 30. For example, the analysis unit 30 determines whether the dialogue between members and between system members in the dialogue thread 22 and the extended dialogue thread 24 is a Type 1 information exchange that satisfies predetermined conditions, or a Type 2 information exchange that does not satisfy predetermined conditions. Then, when generating knowledge, the analysis unit 30 can associate the dialogue information with information on whether the dialogue information is a Type 1 or Type 2 information exchange, and structure the dialogue information.

[0047] For example, the analysis unit 30 can set the aforementioned predetermined condition that the agenda (purpose) of a conversation is achieved through a series of conversations in a dialogue thread 22, etc. This allows the dialogue information accumulation and sharing system 10 to collect data as knowledge on whether each communication in the dialogue unit 20 achieved its purpose, and members can provide this data to the system administrator as needed (see Knowledge 003 in Figure 4 and Knowledge 006 in Figure 5). Furthermore, members and system administrators can obtain an easily understandable indicator of how efficient group communication is from this knowledge.

[0048] Furthermore, when the analysis unit 30 detects a Type 2 information exchange (a conversation that fails to achieve its objective) from the dialogue information, it uses the chat response unit 21 to propose to the members of the dialogue thread 22 that an extended dialogue thread 24 be created, which is a dialogue means for changing the Type 2 information exchange into a Type 1 information exchange (a conversation that achieves its objective). In addition, the analysis unit 30, after obtaining approval from the members, causes the dialogue unit 20 to generate the extended dialogue thread 24 (see Figure 5).

[0049] Furthermore, the analysis unit 30 analyzes the dialogue information between members and between system members in the extended dialogue thread 24, extracts progress information regarding the progress of the task of changing Type 2 information exchange (conversation without achieving the objective) to Type 1 information exchange (conversation with the objective achieved), structures the generation purpose of the extended dialogue thread 24 and the progress information 23c in relation to each other, and generates knowledge (see Figure 5).

[0050] Furthermore, the additional member proposal unit 36 ​​of the analysis unit 30 can refer to the personnel label data (see Figures 4 and 5) and propose members with skills useful for transforming Type 2 information exchange (conversations that fail to achieve their objectives) into Type 1 information exchange (conversations that achieve their objectives) as additional members.

[0051] The following provides a more detailed explanation of the dialogue information aggregation and sharing system 10, but the dialogue information aggregation and sharing system 10 is not limited to those shown in these examples.

[0052] Figure 3 is a flowchart showing an example of the processing performed in the dialogue unit 20 and analysis unit 30 of the dialogue information accumulation and sharing system 10 shown in Figures 1 and 2.

[0053] In step S001 of Figure 3, the dialogue information aggregation and sharing system 10 performs preliminary preparations for the conversation by the dialogue unit 20. Specifically, it generates a dialogue thread 22 and informs members participating in the dialogue thread 22 of the location (URL), etc., of the dialogue thread 22.

[0054] In step S002 of Figure 3, the conversations between members and between system members in the conversation thread 22 are disclosed. The analysis unit 30 of the conversation information aggregation and sharing system 10 analyzes the conversation information between members in the conversation thread 22 and extracts predetermined information 23d such as the purpose (agenda) of the conversation and the start and end points. If the conversation information aggregation and sharing system 10 cannot extract the purpose of the conversation even after analyzing the conversation information, it can inform the members of the conversation thread 22 via the chat response unit 21 to have a conversation that clarifies the purpose (agenda) of this conversation.

[0055] The analysis unit 30 generates information about the purpose of the conversation, extracted from dialogue information, etc., as knowledge and stores it in the memory unit 40.

[0056] In step S003, the analysis unit 30 analyzes the dialogue information 23 between members in the dialogue thread 22 and monitors whether conversations deviating from the objective (agenda, knowledge generated in step S002) are taking place (a type of predetermined information 23d). Examples of conversations deviating from the agenda include conversations unrelated to the agenda or conversations without a specific purpose.

[0057] If the analysis unit 30 extracts a conversation that deviates from the purpose in step S003, the dialogue information collection and sharing system 10 can inform the members of the dialogue thread 22 via the chat response unit 21 to continue the conversation in line with the purpose of the conversation (step S004). For example, if the analysis unit 30 extracts a conversation that deviates from the purpose, the dialogue information collection and sharing system 10 can inform the members via the chat response unit 21 that "Now is not the time" or "Let's talk about that later."

[0058] However, even if the dialogue information collection and sharing system 10 extracts a conversation that deviates from the purpose, it is not always necessary to tell the members to stop it. For example, under certain conditions, the dialogue information collection and sharing system 10 may allow conversations that deviate from the purpose to occur.

[0059] In step S005, the analysis unit 30 (particularly the thread management unit 34) analyzes the dialogue information between members in the dialogue thread 22, extracts and determines the conclusions and endpoints (a type of predetermined information 23d) of the series of conversations that took place in the dialogue thread 22, and generates this as knowledge. The analysis unit 30 can also generate minutes and summaries of the series of conversations that took place in the dialogue thread 22 as knowledge. The storage unit 40 can store knowledge regarding the dialogue between members in the dialogue thread 22, including the purpose of the conversation, the starting point, the ending point, and the conclusions.

[0060] In step S006, the analysis unit 30 analyzes the dialogue information between members in the dialogue thread 22 and determines whether the agenda (objective) of the series of conversations was achieved (objective-achieving conversation 23a (a type of first-type information exchange)). Examples of conversations where the objective was not achieved 23b (a type of second-type information exchange) include those where the objective was not achieved but countermeasures have been decided, and those where the objective was not achieved and countermeasures are unknown.

[0061] If the system determines in step S006 that the purpose of the conversation has not been achieved, the dialogue information aggregation and sharing system 10 can inform the members of the dialogue thread 22 via the chat response unit 21 to perform additional tasks (step S007). For example, if the purpose of the conversation has not been achieved, the dialogue information aggregation and sharing system 10 can, via the chat response unit 21, suggest to the members the schedule for the next meeting or suggest creating an extended dialogue thread with additional members.

[0062] Figure 4 is a conceptual diagram showing an example of dialogue information 23 from a dialogue thread 22 conducted in the dialogue information accumulation and sharing system 10 shown in Figures 1 and 2, as well as knowledge and personnel label data generated by the analysis unit 30 and the personnel label data generation unit 50.

[0063] The left side of Figure 4 shows an example of dialogue information 23 from the dialogue thread 22. In the example shown in Figure 4, a conversation takes place in the dialogue thread 22 in which member A (using terminal 71 in Figure 1) asks a question and member B (using terminal 72 in Figure 1) answers it. The right side of Figure 4 shows the knowledge (knowledge 001, 002, 003) generated by the analysis unit 30 and the personnel label data (skills 101, 102) generated by the personnel label data generation unit 50 from the dialogue information 23 shown in the left side of Figure 4.

[0064] In the example shown in Figure 4, the analysis unit 30 determined that the purpose of the conversation (knowledge 001) extracted from member A's statement was achieved by member B's response (knowledge 003).

[0065] Figure 5 is a conceptual diagram showing another example of the dialogue information of the dialogue thread 22 that takes place in the dialogue information accumulation and sharing system 10 shown in Figures 1 and 2, as well as the knowledge and personnel label data generated by the analysis unit 30 and the personnel label data generation unit 50.

[0066] The left side of Figure 5 shows another example of dialogue information from dialogue thread 22. In the example shown in Figure 5, it can be understood from the dialogue information of dialogue thread 22 that member A asked a question, member B answered it, and then the system (chat response unit 21) intervened between the two and made a suggestion. The right side of Figure 5 shows the knowledge (knowledge 004, 005, 006) generated by the analysis unit 30 from the dialogue information shown in the left side of Figure 5, and that the analysis unit 30 read the personnel label data and suggested additional members.

[0067] In the example shown in Figure 5, the analysis unit 30 determined that the purpose of the conversation (knowledge 004) extracted from member A's statements had not been achieved (knowledge 006), and proposed to the members of dialogue thread 22 (member A, member B) that they carry out a new task (a manual creation meeting in the extended dialogue thread 24) to achieve the purpose of the conversation. [Explanation of Symbols]

[0068] 10…Dialogue Information Collection and Sharing System 20…Dialogue Section 21... Chat Response Department 22... Dialogue thread 23…Dialogue Information 23b...Conversation that fails to achieve its objective 23c… Progress report 23d... Prescribed information 24... Extended Dialogue Thread 30…Analysis Department 32...Knowledge Generation Department 34...Thread Management Department 36…Additional Member Proposal Department 40...Storage section 50... Personnel Label Data Generation Department 60…Human Resources Data Storage Unit 71-74… Terminals

Claims

1. A dialogue unit provides dialogue threads to terminals used by members belonging to a predetermined group, or enables dialogue between members and between system members using a large-scale language model via information transmission and information acquisition means installed in the space where the members reside, and acquires dialogue information between the members and between the system members. An analysis unit analyzes the aforementioned dialogue information and generates structured knowledge from the aforementioned dialogue information, It has a storage unit for storing the aforementioned knowledge, The dialogue unit is a dialogue information aggregation, structuring, and sharing AI system that, in a dialogue between system members, reads the knowledge from the storage unit and conveys the content of the knowledge to the members. The analysis unit determines whether the dialogue between the members and between the system members is a Type 1 information exchange, which is an information exchange that corresponds to a conversation that achieves a predetermined objective and satisfies predetermined conditions, or a Type 2 information exchange, which is an information exchange that corresponds to a conversation that fails to achieve a predetermined objective and does not satisfy predetermined conditions. When generating the knowledge, the analysis unit associates the dialogue information with information indicating whether the dialogue information is the result of a first-type information exchange or a second-type information exchange, and structures the dialogue information. The dialogue unit, in cooperation with the analysis unit, refers to the knowledge regarding the second type of information exchange generated by the analysis unit by structuring the dialogue information, and can interrupt the dialogue between members at the moment the analysis unit detects the second type of information exchange in the dialogue between members in the dialogue unit. This is a dialogue information aggregation, structuring, and sharing AI system.

2. The AI ​​system for accumulating, structuring, and sharing dialogue information according to claim 1, wherein the analysis unit, when generating the knowledge, associates the dialogue information with the member who initiated the dialogue information and structures the dialogue information.

3. The AI ​​system for accumulating, structuring, and sharing dialogue information according to claim 1, wherein the analysis unit extracts and classifies predetermined information from the dialogue information when generating the knowledge, and structures the dialogue information.

4. A dialogue unit provides dialogue threads to terminals used by members belonging to a predetermined group, or enables dialogue between members and between system members using a large-scale language model via information transmission and information acquisition means installed in the space where the members reside, and acquires dialogue information between the members and between the system members. An analysis unit analyzes the aforementioned dialogue information and generates structured knowledge from the aforementioned dialogue information, It has a storage unit for storing the aforementioned knowledge, The dialogue unit is a dialogue information aggregation, structuring, and sharing AI system that, in a dialogue between system members, reads the knowledge from the storage unit and conveys the content of the knowledge to the members. The analysis unit determines whether the dialogue between the members and between the system members is a Type 1 information exchange, which is an information exchange that corresponds to a conversation that achieves a predetermined objective and satisfies predetermined conditions, or a Type 2 information exchange, which is an information exchange that corresponds to a conversation that fails to achieve a predetermined objective and does not satisfy predetermined conditions. When generating the knowledge, the analysis unit associates the dialogue information with information indicating whether the dialogue information is the result of a first-type information exchange or a second-type information exchange, and structures the dialogue information. When the analysis unit detects the second type of information exchange from the dialogue information, it proposes to the members of the dialogue, using the dialogue unit, that an extended dialogue thread be provided, which is a dialogue means for changing the second type of information exchange to the first type of information exchange. The analysis unit causes the dialogue unit to generate the extended dialogue thread, When the analysis unit proposes to the members of the dialogue that an extended dialogue thread be established, it proposes that additional members other than the members of the dialogue be added to the dialogue members in the extended dialogue thread. The dialogue unit is an AI system for accumulating, structuring, and sharing dialogue information that generates the extended dialogue thread by adding the additional member to the dialogue members.

5. The AI ​​system for accumulating and structuring dialogue information according to claim 4, wherein the analysis unit analyzes dialogue information between members and between system members in the extended dialogue thread, extracts progress information regarding the progress of the task of changing the second type of information exchange to the first type of information exchange, structures the progress information in relation to the purpose of generating the extended dialogue thread, and generates the knowledge.

6. A talent label data generation unit extracts information about the skills of the member who initiated the dialogue based on the aforementioned dialogue information, and generates talent label data that associates the member with those skills. The AI ​​system for accumulating and structuring dialogue information according to claim 1, further comprising a personnel data storage unit for storing the aforementioned personnel label data.

7. A talent label data generation unit extracts information about the skills of the member who initiated the dialogue based on the aforementioned dialogue information, and generates talent label data that associates the member with those skills. The system further includes a personnel data storage unit that stores the aforementioned personnel label data, The AI ​​system for accumulating and structuring dialogue information according to claim 4, wherein the analysis unit reads the personnel label data from the personnel data storage unit and proposes adding members having the skills useful for changing the second type of information exchange to the first type of information exchange as additional members to the extended dialogue thread.

Images