System for automated management / moderation of meetings

Abstract

System product for automatically (by a computer-implemented bot) led / moderated meetings with 2 or more participants, wherein the bot technically receives and follows specific instructions for leading / moderating meetings, uses a current computer-implemented model (in the sense of artificial intelligence), identifies requests to the model with a tag for the respective person, and considers the previous context (i.e., the previous statements of all persons and the previous answers of the model) during the course of the meeting.

Description

[0001] Recent advances in artificial intelligence allow for lifelike communication between a human and a computer-implemented model via a computer-implemented application. This applies, for example, to GPT-4 and the ChatGPT application from OpenAI. At the same time, there is a growing need in our society for meetings to actually and efficiently lead to results. (Instead of "meetings," terms such as "meetings," "conversations," "telcos," "calls," and similar terms are also used—for example, in the business / official sphere.)

[0002] Hence the idea of ​​having meetings controlled by a computer-implemented application, which in turn accesses a model. (Controlling a meeting is also called "leading" or "moderating".) - The technical problem here is that such applications currently assume that one person is communicating with the model (i.e. not several people), whereas meetings naturally always involve at least two people (here referred to as "participants", "participating persons" or "TP"). It is not a solution for the participants to jointly use a single, simple user interface system, as currently exists, because then the model cannot distinguish who made which statement. It is also not a solution for each participant to use their own user interface and for these interfaces to be completely independent of one another, because then 1.the requests to the model cannot be correlated with each other and 2. the participants can only perceive their own communication.

[0003] Furthermore, a meeting cannot be led / moderated by a single response from a model or by a single response per participant, as a meeting always involves a longer communication process between the participants. This requires a technical solution for both transmitting instructions for the specific design of the leadership / moderation (including control if the meeting has multiple goals and / or topics and / or at least one goal / topic consists of several sub-goals / topics) and providing the model with relevant information from the previous course of the specific meeting.

[0004] All this is technically solved by a computer-implemented system product according to the following description: An environment within a (computer-implemented) communication system (1) is set up and made available to all participants of a specific meeting (6). The participants of this meeting can perceive what each participant says in the meeting. At the same time, this provided environment is separated from other environments (belonging to other meetings) to ensure confidentiality.

[0005] Within the provided environment, a computer-implemented "bot" (2) operates to control the meeting. (It could also be referred to as a "robot," "virtual assistant," "virtual moderator," or similar.) This bot transmits all specified instructions for chairing the meeting (3) (e.g., specifications regarding the decision-making process depending on the decision-making authority of the participants), the first utterance of a participant, and an identifier for this participant (!) to a model (4) (e.g., a pre-trained large language model), which then responds accordingly. The bot outputs this response in the communication system and stores the original utterance of the TP and the model's response in a computer-implemented storage (5). For each subsequent utterance by one of the participants, the bot extracts the previous utterances and answers (i.e., the context) from the storage and also transmits them to the model (together with the instructions and the new utterance).

[0006] If basic rules of communication are not observed by TPs (e.g., if TPs speak inappropriately at the same time), the bot can, if necessary, criticize the TP's input directly (i.e., without involving the model) and ask for a rule-compliant repetition of the statements.

[0007] If the participants in a meeting use a common user input channel (e.g., a shared room microphone), the invention described here is enriched by an additional component called a "separator." This component (1) separates the shared input communication stream received from the participants into individual utterances, so that (2) the individual utterances can be labeled with a marker for each participant. This separation occurs, for example, according to different speech, visual, or audiovisual features of the participants.

[0008] For a successful technical system product, it is also important that two environments provided in (1) at the same time have identical properties. Only then can an intended output be achieved by the technical system product. - The identical properties cannot be reliably achieved if the provision of the environments is done entirely or partially manually. Therefore, the following occurs before a meeting: If the participants of a meeting (or one of the participants as a representative) issue an order (7) for automatic management / moderation of a meeting, a computer-implemented automatism is run through which completely creates and provides the environment in (1) with the predefined properties, with all components and with access for the participants and then transmits the necessary access data to the participants.If desired, the provided environment can then be reused not only for one meeting, but also for subsequent meetings.

[0009] The invention described here achieves, among other advantages, that the aforementioned technical problems are overcome and that none of the participants has to devote their energy to leading a meeting, allowing all participants to fully concentrate on the topics to be discussed in the meeting and the meeting objectives. Furthermore, the invention presented here prevents those moderating or leading meetings from abusing their implicit power in the meeting for their own benefit. Both of these factors lead to more meetings actually producing results more efficiently. If desired, the invention can also ensure the appropriate level of documentation for meeting results.

[0010] Although the description of the invention describes one or more embodiments, configurations, or aspects, as well as certain variations and modifications, other variations, combinations, and modifications are also within the scope of the invention. It is intended to obtain rights that include alternative embodiments, configurations, or aspects to the extent permissible, including alternative, interchangeable, and / or equivalent structures, functions, ranges, or features to those claimed, whether or not such alternative, interchangeable, and / or equivalent structures, functions, ranges, or features are disclosed herein.

Claims

[1] System product for automatically (by a computer-implemented bot) led / moderated meetings with 2 or more participants, whereby the bot technically receives and observes specific instructions for the management / moderation of meetings, uses a current computer-implemented model (in the sense of artificial intelligence), provides requests to the model with a label for the respective person and whereby the previous context (ie the previous statements of all persons and the previous answers of the model) is taken into account during the course of the meeting [2] System product as in No. 1 for the embodiment that individual statements of the participating persons (e.g. in the case of improper simultaneous speaking) are subjected to special treatment, such as a request for repetition (The scope of the entire invention is not, however, limited to this embodiment). [3] System product as in No. 1 or 2 for the embodiment in which a common input communication stream of several participating individuals is separated into the utterances of each individual before the requests are sent to the model. The separation is performed, for example, according to different speech features, visual features, or audiovisual features of the TP. (However, the scope of the entire invention is not limited to this embodiment.) [4] System product as in No. 1, 2 or 3 for the embodiment that a Large Language Model is used (The scope of the entire invention is not limited to this embodiment, however.) [5] System product as in No. 1, 2, 3 or 4 for the embodiment that the instructions technically transmitted to the bot are specific for the discussion method “brainstorming” (The scope of the entire invention is, however, not limited to this embodiment.) [6] System product as in Nos. 1, 2, 3, 4 or 5 for the embodiment that serves as context a predetermined but configurable set of previous statements and answers (The scope of the entire invention is, however, not limited to this embodiment.) [7] System product as in Nos. 1, 2, 3, 4 or 5 for the embodiment that serves as context a summary of all previous statements and answers in the discussion (The scope of the entire invention is not, however, limited to this embodiment.) [8] System product as in Nos. 1, 2, 3, 4, 5, 6 or 7 with the additional option that the context can be deleted if necessary, so that the subsequent answers of the model are no longer influenced by the statements made by the participants up to that point and the answers received by the model up to that point (The scope of the entire invention is not, however, limited to this variant). [9] System product for the automatic, computer-implemented provision of a confined environment within the system product according to Nos. 1, 2, 3, 4, 5, 6, 7 or 8 [10] System product as in No. 9 for the embodiment that the demarcated environment was requested and serves for a single group of people and for their meeting or meetings (The scope of the entire invention, however, is not limited to this embodiment.)

Images