Scaling Laws: Let's Do the Science! Talking Algorithms with Cathy O'Neill

Cathy O’Neil, CEO of ORCAA and author of Weapons of Math Destruction and The Shame Machine, joins Kevin Frazier, AI Innovation and Law Fellow at the University of Texas School of Law and Senior Editor at Lawfare, to explore the promises and limits of algorithmic auditing. The conversation examines what audits actually do in practice, how organizations measure and mitigate bias, and why context—not just code—determines whether an AI system causes harm. O’Neil explains why auditing cannot be reduced to a checklist, where it can meaningfully improve outcomes, and where it risks creating a false sense of security. They also discuss the need for evidence-based AI policy, the challenges of translating ethical concerns into measurable standards, and how regulators should think about auditing as part of broader governance frameworks. Logan Le-Jeffries, a wonderful member of the AI Innovation and Law Program, provided research assistance on this episode.