Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

PTFA: An LLM-based Agent that Facilitates Online Consensus Building through Parallel Thinking

Wen Gu, Zhaoxing Li, Jan Buermann, Jim Dilkes, Dimitris Michailidis, Shinobu Hasegawa, Vahid Yazdanpanah, Sebastian Stein

TL;DR

The paper addresses the difficulty of online consensus building amid diverse stakeholder opinions and limited facilitation expertise. It introduces PTFA, an LLM-based agent that operationalizes Parallel Thinking through the Six Thinking Hats to perform multiple facilitator roles in real time within an online, text-based platform. A pilot study with 48 participants across 16 groups demonstrates PTFA’s potential for idea generation and emotional probing, and yields a new dataset capturing both participant–participant and participant–agent interactions. However, the study also reveals challenges in timing, phase management, and intervention cadence, indicating necessary directions for future work to scale and regulate LLM-facilitated facilitation in multi-user settings.

Abstract

Consensus building is inherently challenging due to the diverse opinions held by stakeholders. Effective facilitation is crucial to support the consensus building process and enable efficient group decision making. However, the effectiveness of facilitation is often constrained by human factors such as limited experience and scalability. In this research, we propose a Parallel Thinking-based Facilitation Agent (PTFA) that facilitates online, text-based consensus building processes.The PTFA automatically collects real-time textual input and leverages large language models (LLMs)to perform all six distinct roles of the well-established Six Thinking Hats technique in parallel thinking.To illustrate the potential of the agent, a pilot study was conducted, demonstrating its capabilities in idea generation, emotional probing, and deeper analysis of idea quality. Additionally, future open research challenges such as optimizing scheduling and managing behaviors in divergent phase are identified. Furthermore, a comprehensive dataset that contains not only the conversational content among the participants but also between the participants and the agent is constructed for future study.

PTFA: An LLM-based Agent that Facilitates Online Consensus Building through Parallel Thinking

TL;DR

The paper addresses the difficulty of online consensus building amid diverse stakeholder opinions and limited facilitation expertise. It introduces PTFA, an LLM-based agent that operationalizes Parallel Thinking through the Six Thinking Hats to perform multiple facilitator roles in real time within an online, text-based platform. A pilot study with 48 participants across 16 groups demonstrates PTFA’s potential for idea generation and emotional probing, and yields a new dataset capturing both participant–participant and participant–agent interactions. However, the study also reveals challenges in timing, phase management, and intervention cadence, indicating necessary directions for future work to scale and regulate LLM-facilitated facilitation in multi-user settings.

Abstract

Consensus building is inherently challenging due to the diverse opinions held by stakeholders. Effective facilitation is crucial to support the consensus building process and enable efficient group decision making. However, the effectiveness of facilitation is often constrained by human factors such as limited experience and scalability. In this research, we propose a Parallel Thinking-based Facilitation Agent (PTFA) that facilitates online, text-based consensus building processes.The PTFA automatically collects real-time textual input and leverages large language models (LLMs)to perform all six distinct roles of the well-established Six Thinking Hats technique in parallel thinking.To illustrate the potential of the agent, a pilot study was conducted, demonstrating its capabilities in idea generation, emotional probing, and deeper analysis of idea quality. Additionally, future open research challenges such as optimizing scheduling and managing behaviors in divergent phase are identified. Furthermore, a comprehensive dataset that contains not only the conversational content among the participants but also between the participants and the agent is constructed for future study.

Paper Structure

This paper contains 14 sections, 7 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. Group Decision Support with Automated Facilitation
  4. Facilitation with the Six Thinking Hats Method
  5. Methodology
  6. User Study
  7. Participants & Study Setting
  8. Results
  9. Survey Responses
  10. Constructive Facilitation Patterns
  11. Problematic Facilitation Patterns
  12. Discussion
  13. Conclusions
  14. Acknowledgments.

Figures (7)

  • Figure 1: The framework of PTFA.
  • Figure 2: Screenshot of online discussion platform
  • Figure 3: The characteristics of the participants in terms of age, gender and English proficiency.
  • Figure 4: Survey responses for the questions around the rating of the facilitator models and the agreement with the resulting consensus. The legend in Figure \ref{['fig:surveyResults:legendAgreement']} shows the possible responses used for the questions in Figure \ref{['fig:surveyResults:facilitatorRating']} and Figure \ref{['fig:surveyResults:consensusAgreement']}. The 50% line divides the bars at 50% of participants.
  • Figure 5: Examples of constructive facilitator contributions
  • ...and 2 more figures