Stanford Computational Antitrust Episode 32: Cary Coglianese on the Use of AI in Public Enforcement
Mar 12, 2025
Cary Coglianese, the Edward B. Shils Professor of Law and Political Science at the University of Pennsylvania, dives into the intriguing world of AI in public enforcement. He discusses the shift in AI regulation from rigid 'guardrails' to more flexible 'leashes' that embrace innovation while ensuring safety. Coglianese addresses the balance of AI's efficiencies against ethical standards, highlighting the necessity of human oversight and the implications of recent regulatory developments, including the EU's Digital Markets Act.
AI Snips
Chapters
Books
Transcript
Episode notes
AI Regulation Approaches
- AI regulation approaches range from strict "guardrails" to flexible "leashes".
- "Guardrails" are fixed rules, while "leashes" allow for dynamic adjustments.
Leashes vs. Guardrails
- Traditional AI was task-specific, but generative AI is multifunctional, changing the regulatory landscape.
- "Guardrails" are unsuitable for AI's dynamic nature; "leashes" offer flexibility with human oversight.
AI Safety and Geopolitics
- The Paris AI Summit shifted focus from AI safety, reflecting geopolitical competition.
- Safety remains crucial for AI development to avoid disastrous consequences and backlash regulation.