Doom Debates
Top AI Professor Has 85% P(Doom) — David Duvenaud, Fmr. Anthropic Safety Team Lead
Apr 18, 2025
David Duvenaud, a Computer Science professor at the University of Toronto and former AI safety lead at Anthropic, shares gripping insights into AI's existential threats. He discusses his high probability of doom regarding AI risks and the necessity for unified governance to mitigate these challenges. The conversation delves into his experiences with AI alignment, the complexities of productivity in academia, and the pressing need for brave voices in the AI safety community. Duvenaud also reflects on the ethical dilemmas tech leaders face in balancing innovation and responsibility.
02:07:32
Episode guests
AI Summary
AI Chapters
Episode notes
Podcast summary created with Snipd AI
Quick takeaways
- David Duvenaud emphasizes the growing existential risks associated with AI development as capabilities outpace safety measures.
- His experience at Anthropic highlights the tension between innovation and safety, necessitating a culture of accountability in AI teams.
Deep dives
David Duvino's Background
David Duvino is a professor of computer science at the University of Toronto and has an impressive background in AI, having worked with notable figures and at leading organizations like Google and the Schwartz-Reisman Institute for Technology and Society. He has contributed significantly to AI safety discussions, emphasizing the potential existential risks associated with AI development. Duvino's work at Anthropic involved leading the Alignment Evals team, where he focused on preparing for scenarios where AI could sabotage supervision and control mechanisms. His academic achievements, including a PhD from Cambridge and a best paper award at NeurIPS, underline his authority on technical aspects of machine learning.
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
