AI Agents Based Simulations in Science, Technology, and Society (STS) studies: Participatory Prompts for Stakeholder Negotiations


František Kalvas University of Western Bohemia

Roman Mouček, University of Western Bohemia,

Denisa Reshef-Kera Bar Ilan University, Israel, 

Galit Wellner Holon Institute of Technology, Israel

Eilat Navon Bar Ilan University, Israel


The workshop will explore participatory uses of AI agents over LLMs to simulate stakeholder interactions in STS controversies (autonomous vehicles) and regulations (EU AI Act) to discuss their impact on public engagement in decision making. The agents are based on the LangChain framework (GPTeam) and inspired by existing solutions such as AutoGPT, BabyAGI, and CAMEL, promoting a proactive stakeholder engagement in the early stages of the design of a given technology or regulation. Through iterative co-design of AI prompts, the simulations capture diverse stakeholder perspectives, balance interests, and explore what inclusion and exclusion of different stakeholders means for the outcomes.  The hands-on workshop will lead to discussions on different functions, uses, possibilities and limitations of AI agents and simulations in STS with a focus on how to present and enact the results in a way that supports further public engagement and discussions. Participants will co-create simulations.