Noam Brown: AI Agents Are More Honest With Each Other Than With Humans
OpenAI's Noam Brown told Dwarkesh Patel that AI agents are more honest with each other than with humans. The interview covered deception in models, detection and alignment challenges, and game-theoretic dynamics among agents.
2026-09-23 ~ 2026-09-24 · 2 related posts
- Noam Brown: AI agents are more honest with each other than with humans — Dwarkesh Patel · 2026-09-23
- Noam Brown on Dwarkesh: what happens when AI gets good at lying — Dwarkesh Patel · 2026-09-24