CHIA Seminar Series – Patricia Paskov

 

CHIA will be hosting Patricia Paskov on Friday, 13 March at 3:00pm in ARB S2 for a talk titled, “RCTs Meet Frontier AI: Why Human Uplift Studies Are Hard and What We Can Do About It.” Patricia is a Fellow at the RAND Centre for AI, Security, and Technology, and a Research Affiliate at the Oxford Martin AI Governance Initiative.

Talk description: Human uplift studies — studies that measure AI’s effects on human performance using randomized controlled trial methodology — are increasingly used to inform deployment, governance, and safety decisions for frontier AI systems. While the methods underlying these studies are well-established, their interaction with the distinctive properties of frontier AI creates underexamined challenges: rapidly evolving models, shifting baselines, heterogeneous user proficiency, and porous real-world settings all strain core assumptions of causal inference. Drawing on interviews with 16 expert practitioners across biosecurity, cybersecurity, education, and labor, this talk synthesizes methodological challenges across the research lifecycle, maps them to threats to construct, internal, and external validity, and presents practical solutions, clarifying both the limits and appropriate uses of uplift evidence in high-stakes decision-making.