This session explores how AI can function as deliberative infrastructure in high-stakes political environments rather than as a tool for persuasion or automation. Drawing on CRISP’s 18 years of experience in crisis simulations and recent deployments of AI-supported discussion systems in conflict and governance contexts, the talk presents concrete lessons from using AI as a bounded advisor, sparring partner, and simulation component.

The focus lies on trust, legitimacy, perspective change, and facilitation: what design choices matter when AI is embedded in collective political processes under pressure, and what democratic and political tech actors can learn from crisis and conflict settings.