Test, Learn, Regulate – How Experimental Approaches Are Shaping AI Governance [GFIN]

This roundtable will discuss the evolution of regulatory approaches in response to rapidly advancing AI technologies. The event aims to gather global regulators, academics, and industry experts to explore adaptive, collaborative regulatory methods for AI governance.
Purpose and participants: The roundtable intends to examine how regulatory practices must evolve as AI systems outpace traditional rulemaking, involving AI Compass participants and broader GFTN network members and academics.
Key objectives: It seeks to clarify regulatory challenges, showcase test-and-learn models like AI sandboxes and tech sprints, address supervisory issues such as explainability and bias, identify cross-border cooperation opportunities, and highlight near-term shifts in AI governance over the next three years.
Discussion themes: The session will explore how regulators are building capabilities through experimental methods, tackle supervisory considerations including model drift and third-party dependencies, and discuss proportionality and assurance ecosystems to manage risks without hindering innovation. 
Cross-border collaboration and future outlook: The roundtable will consider realistic international cooperation such as shared research and sandbox exchanges, conclude with rapid-fire insights on necessary regulatory shifts in the next three years, and reflect on accountability in an AI-driven future.