Four Approaches to Developing Autonomous Facilitator Agent for Online and Face-to-Face Public Debate

Recently, the importance of public debate is increasing both globally and locally for addressing sustainability problems such as pandemics, climate change, and economic crisis. To support such public debate, software agents need to be developed to facilitate discussions, for example, to recommend relevant information by detecting stagnation and flaming in online public debate, to invite debate participants from SNS, or to record face-to-face public debates. In this study, we prototyped four software agents for facilitation: (1) an agent for detecting stagnation and flaming while quantifying the degree of discussion progress in a Web-based debate, (2) an agent for providing relevant information in accordance with the preceding context of a Web-based debate, (3) an agent for finding people who are interested in the content of the discussion and inviting them to a public debate from Twitter, and (4) an agent for recording a face-to-face public debate and supporting users’ reviewing of the debate. In this paper, we overview these four agents and evaluation experiments and present the feedback from the participants in an event organized by Facilitation Association of Japan.