Henry Papadatos

Background

About Henry Papadatos

My focus lies in addressing the growing gap between rapid AI advancement and our ability to manage its associated risks. I am the Managing Director of SaferAI where I work on both AI governance and technical solutions to improve the risk management for frontier AI systems. Governance: I contribute to shaping international AI policy through multiple channels. I was a member of one of the expert working groups for the EU AI Act's Codes of Practice, focusing on risk taxonomy and assessment for general-purpose AI models. I participated in the OECD task force that drafted the G7 Hiroshima AI Process reporting framework and recently joined the OECD expert group on AI Risk & Accountability. My advocacy for international AI safety standards was featured in a TIME op-ed ahead of the Paris AI Action Summit. Technical Research: My technical work centers on developing practical tools for AI risk management. I created an AI risk management ratings system for developers (featured twice in TIME, the Guardian, Euractiv and others) and published a comprehensive framework that integrates current AI practices with proven strategies from other industries. My current research focuses on improving risk modeling and quantitative assessment for AI-enabled cyber risks, including recent work connecting empirical AI capability measures to concrete real-world harm estimates. Prior to joining SaferAI, I conducted technical research on large language model alignment at the Center for Human-Compatible AI at UC Berkeley.