HUMAN-ARTIFICIAL INTELLIGENCE HYBRID SWARM INTELLIGENCE AND COLLECTIVE DECISION-MAKING SYSTEM

The hybrid swarm intelligence system integrates human and AI participants with dynamic reliability and ethical oversight, addressing the limitations of existing systems by enhancing real-time decision-making and compliance monitoring.

TR202508032A3Pending Publication Date: 2026-10-21TURKIYE GARANTI BANKASI ANONIM SIRKETI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
TR202508032
Authority / Receiving Office
TR · TR
Patent Type
Applications
Current Assignee / Owner
Filing Date
2025-06-17
Publication Date
2026-10-21

AI Technical Summary

Technical Problem

Existing swarm intelligence systems lack proactive participation and deep integration of artificial intelligence agents with humans, insufficient reliability mechanisms, and fail to address ethical and regulatory compliance in collective decision-making processes, particularly in regulated sectors like banking.

Method used

A hybrid swarm intelligence system combining human and AI participants with dynamic, self-improving reliability through multi-layered conformal estimation, ethical oversight, and continuous learning, utilizing Big Language, Vision, Speech, and World Models to enhance real-time decision-making and compliance monitoring.

Benefits of technology

Enhances the reliability and ethical compliance of collective decision-making by dynamically adjusting AI agent contributions, ensuring consistent and contextually relevant participation, and actively monitoring for regulatory violations.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

This invention relates to a system (1) that enables dynamic, self-improving and multi-layered reliability-focused collective decision-making through participants consisting of humans and artificial intelligence.
Need to check novelty before this filing date? Find Prior Art

Description

1 TARIFF HUMAN-ARTIFICIAL INTELLIGENCE HYBRID SWARM INTELLIGENCE AND COLLECTIVE DECISION-MAKING ALMA SYSTEM Technical Area 5 This invention will be made possible through a collaboration between human and artificial intelligence participants. Dynamic, self-improving, and multi-layered reliability-focused collective decision-making. It is related to a system that provides. Previous Technique In the current state of the art, increasing the effectiveness of collective decision-making processes. This quest has led to the development of various technological and methodological approaches. This has led to the development of the concept of swarm intelligence, which has been embodied in some developed systems. This 15 systems that simulate the collective intelligence of human groups, usually artificial intelligence by strengthening it through moderator or simple machine predictions, the truth It aims to reach a common decision through timely interactions. In some systems, artificial intelligence is used. Intelligence input is typically based on intuitive methods or human proxy data. It is defined as "machine prediction" based on trained models, and this prediction is 20 "Agent intent" affects a marker with which participants interact. They are converted into "vectors". Some studies weight participant responses by and adaptive trust that calculates group intent by learning user trust scores. It offers a calibration mechanism. Artificial intelligence, especially Big Tongue. Their models are equipped with natural language processing, complex data analysis, and reasoning capabilities. It is widely used in decision support systems. However, these uses... It usually provides support to the individual user or uses artificial intelligence for passive analysis. It tends to position itself as an intermediary. Multi-agent systems have multiple autonomous agents cooperate or compete to achieve specific goals It offers frameworks. Some systems are specialized for complex tasks such as research synthesis. It uses multiple AI agents with different roles; however, such systems 2 rather than real-time, interactive collective decision-making processes, It has a pipeline-based structure for automated report generation. Existing Despite these advances in technology, the following significant shortcomings and unmet demands remain: The needs remain: Artificial intelligence in existing swarm intelligence systems, They generally assume limited roles. Multiple, heterogeneous (Large Language Models, 5 Big Image Models, Big Speech Models, and the Big World models) and equipped with different, dynamically adaptable "personas" Unanimously, the agent was a proactive and sophisticated participant on an equal footing with the people. Its deep integration into the herd is insufficient. Artificial intelligence, only Instead of responding only upon request, all participants in the herd (human and other 10) artificial intelligence) to understand complex topics and data sets more deeply. To provide this, we proactively utilize multimodal (text, visual data interpretations, (scenarios) producing, synthesizing and continuously improving content This has not been addressed sufficiently. The contributions of the participants (human or artificial intelligence) its reliability extends beyond individual model calibration, through multi-layered 15 (personal background consistency, suitability for assigned persona, current discussion) the logical appropriateness of the contribution within the context, peer feedback) based on criteria dynamic evaluation using conformal estimation methods and this assessment modulates the weight of influence in the decision-making process in real-time. This is a new approach. In the current patent literature, a conformal 20 of this depth is unprecedented. The prediction application is not specifically defined. The system only applies to individual users. not only the capabilities of artificial intelligence agents, but also swarm interaction. protocols, AI persona assignment strategies, and general decision-making its mechanisms are derived from past herd experiences and "the distillation of collective wisdom" by learning through, it has developed a comprehensive 25 autonomous and integrated system. An evolutionary cycle is missing. Especially in highly regulated sectors like banking. ethical principles and legal regulations during the decision-making process in established sectors and actively monitor institutional policies, flag potential violations, and even a special "Ethics and The existence and functionality of the "Compliance Monitoring AI Agent" is sufficient. 30 Undefined. 3 Therefore, considering the studies and shortcomings in the current technique... when present, human participants and multiple, diverse personas, By synergistically combining proactive content-generating AI agents, decision-making... dynamically managing its reliability with multi-layered conformal estimation, continuous 5 by learning, it autonomously develops both its agents and system strategies, and with a hybrid swarm intelligence that has an integrated ethical oversight mechanism dynamic, self-improving and multi-layered reliability-focused collective It appears that a system providing a decision-making mechanism is needed. United States Regulation US11151460B2, which is included in the known state of the art. The patent document describes the collective intelligence of human communities as an intelligent choice. software that uses optimized subpopulations by organizing them through a process a system that strengthens the accuracy of predictions produced by systems and the method is mentioned. In the invention in question, participants evaluated the event results on 15 The system makes predictions and / or evaluates confidence in its predictions. the participants' responses and the participants' responses as a whole of the population It assigns an outlier score to each participant based on their relationship to the predictions. Participants can then be selected from the population based on their extreme point scores. Another document included in the known state of the art is US11941239B2. In the United States patent document number [number], two possible future events are listed. collective of networked human groups making collaborative predictions resulting from The text describes a system and method that enhances intelligence. Brief Description of the Invention The goal of this invention is to create a group of participants consisting of humans and artificial intelligence. through dynamic, self-improving and multi-layered reliability-focused The goal is to implement a system developed to facilitate collective decision-making. 30 4 Detailed Description of the Invention The "Human-Artificial Intelligence Hybrid Swarm" was developed to achieve the purpose of this invention. The "Intelligence and Collective Decision-Making System" is shown in the attached diagram; Figure 1 shows a schematic view of the system that is the subject of the invention. The parts shown in the figure are individually numbered, and these numbers correspond to... The corresponding answers are given below. 1. System 2. Electronic Devices 3. Interface 4. Server A dynamic, self-contained environment created through participants consisting of humans and artificial intelligence. Developing and enabling multi-layered, reliability-focused collective decision-making. The system in question, developed for the purpose of invention (1); - exchanging data using any remote communication protocol at least one electronic device configured to perform (2), 20 - to be run on electronic device (2), user input to enable them to do so, the details of the problem to be discussed or solved to ensure that it is defined in some way, the initial data to enable loading, to enable setting session goals and enabling the session to be initiated, 25 by artificial intelligence. the rich ingredients produced / synthesized in an interactive way display and be centralized or distributed, intelligent filtering, It has the ability to summarize and deliver personalized content. at least one interface structured to be (3), - using any remote communication protocol electronic device (2) 30 to establish a connection and through this connection, an electronic device (2) accessing the interface (3) running on the interface (3) Accessing user information that logs in to face (3), various Major Language Models, Large Vision Models, Large Speech Models, and Artificial intelligence agents in the form of Big World Models or Running combinations of these, each AI agent, 5 Specific knowledge, skills, and requirements based on context analysis and session objectives. identities involving different communication styles or cognitive approaches to equip, to automatically assign these identities during the session, to change or activate, which type of herd at which stage 10 that it needs more expertise or approach by evaluating who to optimize, with artificial intelligence agents The content includes live discussions, and interaction between humans and other artificial intelligence. with feedback from participants and emerging new to constantly update with information and make it more understandable, actionable to make it convertible, the resulting models and analyses 15 by using successful problem-solving strategies, effective arguments patterns, innovative combinations of ideas, or specific problems "Collective wisdom" for determining the most suitable AI identity sets for different types. abstract in this way, present on the interface (3), all data secure, integrated, accessible, centralized or distributed 20 Store, manage, and back up, in real-time, low-cost. to include a communication network that provides delayed but secure data communication and to ensure the highest level of data privacy and security It includes at least one configured server (4). The electronic device (2) in the system (1) which is the subject of the invention, any remote communication exchanging data using the protocol, tablet and / or portable to be a smart device in the form of a computer, to run the interface (3) on it and any remote communication protocol included in the known state of the art It is configured to connect to the server (4) using 30 6 The interface (3) in the system (1) which is the subject of the invention, on the electronic device (2) to be executed, input from users in the form of human participants and system administrators. to enable them to do so, a detailed explanation of the problem to be discussed or solved to enable identification, in the form of documents, data sets, and visual materials To ensure the relevant initial data is loaded, session objectives 5 to enable identification and initiate the session It is structured. The interface (3) allows human participants to interact with texts, sounds, with structured votes or graphical tools and autonomous artificial intelligence agents produced as; text, analysis, script or predefined interaction He expressed his opinions, presented his arguments, and answered questions through the protocols. 10 the rich person who asked, participated in the voting, and was generated / synthesized by artificial intelligence. Displaying content interactively and being centralized or distributed is excessive. Smart filtering, summarizing, and personalized options to prevent information overload. It is structured to have content delivery skills. The server (4) in the system in question (1) is located in the known state of the art. using any remote communication protocol with electronic devices (2) to establish a connection and through this established connection on the electronic device (2) The server (4) is configured to access the interface (3) being executed. Accessing user information via the interface (3), interface (3) 20 through; access to information about the problem to be discussed or solved, documents, Datasets, access to relevant initial data in the form of visual materials, swarm to access the target information defined by the session, the entered problem definition and By analyzing the initial data, the key topics, key concepts, and By identifying areas of potential complexity, artificial intelligence can, through this analysis, 25 a process for selecting agent personas and determining content production strategies It is configured to generate input. The server (4) is configured to generate input for human participants. managing identity verification, areas of expertise, roles, and authorizations. It is structured to define the Server (4), various Big Language Models, Large Vision Models, Large Speech Models, and Large World Models 30 running artificial intelligence agents in this form, or combinations thereof, each 7 The AI ​​agent provides specific information based on context analysis and the session's objectives. skills and identities involving different communication styles or cognitive approaches to equip, automatically assign, and change these identities during the session or to activate, at what stage of the herd, what type of expertise or approach 5 people to optimize by assessing who needs it most. is configured. The server (4) is text-based through Big Language Models. in-depth analyses, comprehensive summaries, dialogue initiators, alternative solutions Producing recommendations and draft reports, presented through Big Vision Models textual insights, pattern identifications, or anomalies by analyzing visual data. 10 for presenting findings, visualizing complex concepts or data relationships Generating textual definitions and instructions through Big Speech Models Converting audio input to text, summarizing key discussion points in audio format. presenting or narrating textual content with different accents to increase accessibility and increasing comprehensibility, complex, long-term through Big World Models and by simulating multivariate scenarios, different strategies or events 15 potential outcomes and probabilities structured textual reports, to present in the form of comparative analyses or decision trees It is structured. The server (4) uses artificial intelligence agents to process the content; live through discussions and feedback from human and other AI participants and to constantly update it with emerging new information and make it more understandable, 20 It is structured to make it actionable. Server (4), intermediate over one hundred (3) human participants; with texts, sounds, structured votes or graphical tools and AI agents autonomously generating text, analysis, scenario or through predefined interaction protocols 25 participation and rich content generated / synthesized by artificial intelligence to enable interactive display and the central (3) interface or Intelligent filtering to ensure distributed information and prevent information overload. Summarizing and personalized content delivery capabilities on the interface (3) It is configured to provide. The server (4) is configured to provide 30 as specified on the interface (3). the opinions of human and artificial intelligence participants on the problem or sub-questions, 8 It is configured to collect preferences or predictions. Server (4), the reliability of the real-time contribution of each human and artificial intelligence participant and to assess consistency, each participant's current contribution and interaction It is configured to save and organize its patterns. Server (4), 5 of each participant's past decision-making patterns and contributions its accuracy / impact, the consistency of the views it expresses, the assigned artificial intelligence recording behaviors consistent with their identities and previous conformal predictive scores It is configured to keep under control. The server (4) is configured to keep under control the current status of a participant. the reliability of its contribution, statistically robust conformal estimation Based on these principles, the participant's current contribution, their own past statements, 10 taking into account how consistent it is with decisions and established behavioral patterns It is configured to calculate by taking into account the current information of a participant. The server (4) the reliability of its contribution, statistically robust conformal estimation Based on these principles, the current contribution or communication of an artificial intelligence agent style, expected behavior of current assigned or elected identity 15 to calculate by taking into account how well it matches its specifications is structured. The server (4) is configured to show a participant's current contribution. its reliability is based on statistically sound conformal estimation principles. Based on this, the contribution fits into the current context of the ongoing debate, previously presented valid evidence, stated problem objectives and logical inference 20 to calculate by taking into account how well it complies with the rules is structured. The server (4) is configured to show a participant's current contribution. its reliability is based on statistically sound conformal estimation principles. based on the immediate response of other participants (human or artificial intelligence) to this contribution. It is structured to calculate by taking into account the analysis of responses. Server 25 (4), combining multilayer conformal prediction scores, voting, scoring or the current contribution of each participant in the collective convergence process It is configured to dynamically adjust its weight or impact. The server (4) received an unexpectedly low signal from an AI agent. Contributions with conformity (high nonconformity score) are a potential 30 model hallucination, lack of information, out-of-identity behavior, or a systemic one 9 Automatically flagging it as an anomaly and requiring further detailed examination or human intervention. It is configured to transmit to the interface (3) for intervention. Server (4), all interaction data from completed sessions and the final decisions of these sessions to thoroughly analyze the results, successful and unsuccessful decision-making patterns, effective or ineffective communication strategies, efficient problem-solving 5 ways, resulting synergies or negative forms of groupthink It is structured to model the dynamics. The server (4) obtains By using models and analyses, we can identify successful problem-solving strategies and effective methods. argument patterns, innovative combinations of ideas, or specific types of problems Abstracting the most suitable AI identity sets as "collective wisdom", 10 It is configured to present on the interface (3). The server (4) is an individual artificial Continuously monitoring and evaluating the performance of artificial intelligence agents Updating the basic models of intelligence agents through fine-tuning, request optimizing engineering strategies, their access to new information sources to provide or to teach new skills through special reinforcement learning or 15 to initiate meta-learning cycles, to optimize working principles It is structured. The server (4) hosts all discussions, generated and shared. the contents, proposed decisions and decision-making processes, predefined and a constantly updated set of ethical codes, legal regulatory directives, customer privacy, fair lending laws, anti-discrimination provisions, and internal 20 Monitoring policies in real time, potential violations, and ethics When it identifies off-trends, biased statements, or risks of inconsistency, Generating warnings based on severity level, temporarily halting the spread of relevant content. to limit, suggest alternative ethical solutions, or guide human decision-making processes. to mark / veto for submission to review, reliability data also 25 to identify suspicious patterns and provide ethical learning inputs by taking these into consideration. It is structured in such a way. The server (4) displays the outcome of the session, the final decisions taken, The fundamental arguments and justifications leading to these decisions are analyzed by artificial intelligence. The significant and effective content produced assesses participants' conformal prediction and reliability. 30 to produce a comprehensive report that includes, is structured and understandable, and these reports, for auditing, accountability, information sharing, and future analysis. It is configured to enable its use. The server (4) provides all data. to store in a secure, integrated, accessible, centralized or distributed manner, Manage and back up real-time, low-latency, and secure data. to include a communication network that enables communication and to ensure data privacy and security at the highest level. It is structured to provide this at the highest level. Industrial application of the invention In the system that is the subject of the invention (1), 10 through human-artificial intelligence hybrid swarm intelligence. Dynamic, self-improving, and multi-layered reliability-focused collective decision-making. A system is being discussed for this purpose. In this invention, human participants, Great Language Models, Large Vision Models, Large Speech Models, and Large With numerous heterogeneous AI tools, including World Models They are integrated and function as peer participants in an interactive swarm environment. 15 They are seen as having diverse, dynamically adaptable personalities and specialized skills. These AI agents, which possess expertise, can control entire swarms. shared understanding and contextualization among participants (both human and AI) proactively create multimodal content to significantly increase awareness. (e.g., in-depth analyses, contextual summaries, predictive scenarios, visual 20) It produces, improves, and synthesizes textual interpretations of data. The subject of the invention is a problem-based, collectively convergent decision-making mechanism. It uses. A key innovation is that each participant's input (human or artificial intelligence) Its reliability and consistency are based on historical self-consistency, adherence to the assigned personality, and continuity. contextual relevance and discretion of the discourse, peer response analytics 25 multilayer conformal for rigorous evaluation according to criteria such as this. It is a forecasting application. This conformal forecast-driven assessment is used for global decision-making. By dynamically regulating the impact of each contribution in the acquisition process, a robust system is created. It establishes a dynamic trust framework. The system (1) consists of artificial intelligence agents and a swarm. a collective of 30 that analyzed past swarm interactions of operational strategies distilling wisdom and autonomously utilizing artificial intelligence capabilities and systemic protocols. 11 evolving together through an advanced learning engine that improves as It is designed to continuously improve itself. The invention consists of the electronic device (2), interface (3) and server (4) in the system (1) 5. Providing information and consent to the user in accordance with data privacy principles. and operates within the scope of the Personal Data Protection Law (KVKK). It shows. Around these fundamental concepts, the invention focuses on a "Human-Artificial Intelligence Hybrid Swarm". 10 Various applications related to the Collective Decision-Making System (1) with its intelligence It is possible to develop further, and the invention cannot be limited to the examples described here. It is essentially as stated in the requests.

Claims

12 REQUESTS 1. Dynamic, through participants consisting of humans and artificial intelligence, Self-improving and multi-layered reliability-focused collective decision-making developed to provide; 5 - exchanging data using any remote communication protocol at least one electronic device configured to perform (2), - to be run on electronic device (2), user input to enable them to do so, the details of the problem to be discussed or solved to ensure that it is defined in some way, initial data 10 to enable loading, to enable setting session goals and enable the session to be initiated, by artificial intelligence. the rich ingredients produced / synthesized in an interactive way display and be centralized or distributed, intelligent filtering, 15 with summarizing and personalized content delivery capabilities including at least one interface (3) structured as follows; - using any remote communication protocol electronic device (2) to establish a connection and through this connection, an electronic device (2) accessing the interface (3) running on the interface (3) Accessing user information that logs in (3), various Great Language 20 Models, Large Vision Models, Large Speech Models, and Artificial intelligence agents in the form of Big World Models or running combinations of these, each AI agent, Specific knowledge, skills, and requirements based on context analysis and session objectives. 25 with identities that include different communication styles or cognitive approaches to equip, to automatically assign these identities during the session, to change or activate, which type of herd at which stage that it needs more expertise or approach by evaluating who to optimize, with artificial intelligence agents The content includes live discussions, between humans and other artificial intelligence 30 with feedback from participants and emerging new 13 to constantly update with information and make it more understandable, actionable to make the obtained models and analyses convertible by using successful problem-solving strategies, effective arguments patterns, innovative combinations of ideas, or specific problems The most suitable AI identity sets for different types are "collective wisdom" 5 abstract in this way, present on the interface (3), all data secure, integrated, accessible, centralized or distributed Store, manage, and back up, in real-time, low-cost. to include a communication network that provides delayed but secure data communication and to ensure the highest level of data privacy and security 10 a system characterized by having at least one server (4) configured (1).

2. Data exchange using any remote communication protocol. to perform, smart device in the form of tablet and / or portable computer 15 to be, to operate the interface (3) on it and in the known state of the art using any remote communication protocol available to the server (4) characterized by an electronic device (2) configured to establish a connection. A system like the one in Request 1 (1).

3. To be carried out on electronic devices (2), human participants and systems Enabling users, such as administrators, to log in. a detailed definition of the problem to be discussed or solved to provide relevant documents, datasets, and visual materials. Ensuring initial data is loaded, session objectives 25 to enable identification and initiate the session from the above requests characterized by the structured interface (3) a system like any other (1).

4. Human participants can use texts, audio, structured votes, or 30 graphical tools and AI agents autonomously generate text; 14 analysis, scenario or predefined interaction protocols through which he / she expressed his / her opinions, presented his / her arguments, and asked questions, participated in the voting and the rich were generated / synthesized by artificial intelligence. to display content interactively and centrally or distributed to prevent information overload, intelligent filtering, summarizing and 5 to have personalized content delivery capabilities from the above requests characterized by the structured interface (3) a system like any other (1).

5. Using any remote communication protocol, connect electronic device (2) with 10 to establish a connection and through this established connection electronic device (2) server (4) configured to access the interface (3) run on it as in any of the above characterized claims system (1).

6. User information accessed via the interface (3) access, via the interface (3); problem to be discussed or solved access to information in the form of documents, datasets, and visual materials accessing the relevant initial data, the target set by the swarm session. Accessing the information, analyzing the entered problem definition and initial data 20 by outlining the main topics, key concepts and potential By identifying areas of complexity, the artificial intelligence agent can perform this analysis. a process for selecting personas and determining content production strategies characterized by the server (4) configured to generate input a system like any of the above requests (1). 25 7. Managing the authentication of human participants, their areas of expertise, with the server (4) configured to define their roles and authorities as in any of the above characterized claims system (1). 30 8. Various Large Language Models, Large Vision Models, Large Speech Models AI agents in the form of models and Big World Models or running combinations of these, each AI agent, Based on context analysis and session objectives, specific knowledge, skills and different 5 to equip with identities that include communication styles or cognitive approaches, These identities can be automatically assigned, changed, or configured during the session. to activate, at what stage of the herd, what type of expertise or approach by assessing who needs it most, in order to optimize 10 of the above requests characterized by the configured server (4) a system like any other (1).

9. In-depth text-based analyses through major language models, comprehensive summaries, dialogue initiators, alternative solutions, and the report. producing drafts, visual 15 presented through Large Vision Models by analyzing data, textual insights, pattern identifications, or presenting anomaly detections, complex concepts or data relationships Generating textual descriptions and instructions to visualize, Great Converting audio input to text through Speech Models is important. Presenting the points of discussion as audio summaries or textual content 20 increasing accessibility and intelligibility by vocalizing with different accents, Complex, long-term, and multifaceted models are created through Grand World Models. Simulating variable scenarios to explore different strategies or events potential outcomes and probabilities structured textual reports, 25 to present in the form of comparative analyses or decision trees from the above requests characterized by the configured server (4) a system like any other (1).

10. Content generated by AI agents; live discussions, human and other artificial intelligences. with feedback from intelligence participants and the resulting new 30 16 to constantly update with information and make it more understandable, actionable with the server configured to make it convertible (4) as in any of the above characterized claims system (1).

11. Human participants on the interface (3) interact with texts, sounds, with structured votes or graphical tools and artificial intelligence agents autonomously generated; text, analysis, scenario or predefined expressing their ideas and arguments through interaction protocols presenting, asking questions, participating in voting, and AI 10 rich content produced / synthesized by interactive means to enable display and the interface (3) central or distributed to ensure this, intelligent filtering to prevent information overload, Summarizing and personalized content delivery capabilities interface (3) 15 characterized by the server (4) configured to present to it. a system like any of the above requests (1).

12. Human and for the problem or sub-questions defined on the interface (3) artificial intelligence participants' opinions, preferences, or predictions The above 20 is characterized by the server (4) configured to collect. a system like any of the requests (1).

13. The real-time contribution of each human and artificial intelligence participant. to assess its reliability and consistency, each participant's current situation 25 to record and structure its contribution and interaction patterns from the above requests characterized by the configured server (4) a system like any other (1).

14. Each participant's past decision-making patterns and their contributions accuracy / impact, consistency of the views expressed, assigned artificial intelligence 30 17 their behaviors consistent with their identities and previous conformal prediction scores characterized by the server (4) configured to keep records a system like any of the above requests (1).

15. The reliability of a participant's current contribution is statistically determined to be 5 Based on robust conformal estimation principles, the participant's current his contribution, his past statements, decisions and recorded behavior to calculate by taking into account how well it matches the patterns from the above requests characterized by the configured server (4) a system like any of them (1). 10 16. Determine the reliability of a participant's current contribution statistically. Based on robust conformal prediction principles, an artificial intelligence agent their current contribution or communication style, whether assigned or elected How well does it align with the expected behavioral characteristics of your current identity? 15 taking into account that the server (4) is configured to calculate as in any of the above characterized claims system (1).

17. The reliability of a participant's current contribution is statistically determined to be 20 Based on robust conformal forecasting principles, the contribution is ongoing. in the current context of the debate, to valid evidence previously presented, how closely do the defined problem objectives and logical inference rules server configured to calculate taking into account suitability (4) like any of the above claims characterized by 25 system (1).

18. To statistically assess the reliability of a participant's current contribution. Based on sound conformal estimation principles, other participants' this 30 to calculate by taking into account the analysis of the immediate responses to the contribution. 18 from the above requests characterized by the configured server (4) a system like any other (1).

19. Combining multi-layered conformal prediction scores for voting and scoring. or 5 of each participant's current contribution in the collective convergence process configured to dynamically adjust its weight or impact any of the above requests characterized by the server (4) such a system (1).

20. An unexpectedly low 10 from an AI agent. Conformity-based contributions, a potential model hallucination, information the absence of automatic behavior in the form of non-identical behavior or systemic anomaly mark as such and search for further examination or human intervention. characterized by the server (4) configured to transmit to the face (3). a system like any of the above requests (1). 15 21. All interaction data from completed sessions and the final decisions on these matters. to thoroughly analyze the results, successful and unsuccessful decision-making patterns, effective or ineffective communication strategies, efficient problem-solving 20 ways of solving problems, resulting synergies or group thinking with the server (4) configured to model negative dynamics as in any of the above characterized claims system (1).

22. Using the models and analyses obtained, successful problem solving 25 their strategies, effective argumentative patterns, innovative combinations of ideas or the most suitable AI identity sets for specific problem types abstracting as "collective wisdom" and presenting it on the interface (3) from the above requests characterized by the configured server (4) a system like any other (1). 30 19 23. To continuously monitor the performance of individual artificial intelligence agents and to evaluate and fine-tune the basic models of artificial intelligence agents updating through, optimizing requests engineering strategies, to provide them with access to new information resources or new skills 5 special reinforcement learning or meta-learning cycles for their learning to initiate, structured to optimize operating principles any of the above requests characterized by the server (4) such a system (1).

24. All discussions, content produced and shared, proposed decisions, and decision-making processes, a predefined and continuously updated system series ethical guidelines, legal regulatory directives, customer privacy, fairness lending laws, anti-discrimination provisions, and internal policies In terms of real-time monitoring, potential violations, unethical practices 15 When it identifies trends, biased statements, or risks of inconsistency, Generating warnings based on severity level, temporarily preventing the spread of relevant content. to limit as such, to suggest alternative ethical solutions, or to make specific decisions to mark / veto the paths to be submitted to human control, Identifying suspicious patterns by also considering reliability data and 20 with server (4) configured to provide ethical learning inputs as in any of the above characterized claims system (1). The outcome of the 25th session, the final decisions taken, and the underlying principles that led to these decisions. arguments and justifications, significant and effective, generated by artificial intelligence. the content includes participants' conformal estimation and reliability analyses, including the findings of artificial intelligence and a summary of the entire decision-making process, to produce a structured, understandable, and comprehensive report, and these reports, 30 for auditing, accountability, information sharing and future analysis characterized by the server (4) configured to enable its use. a system like any of the above-mentioned requests (1).

26. All data must be stored securely, integrally, accessibly, in a centralized or distributed environment. to store, manage and back up in real-time, low 5 to include a communication network that provides delayed and secure data communication and to ensure the highest level of data privacy and security from the above requests characterized by the configured server (4) a system like any other (1). 15 25