Safety
In an interview in this article for The @guardian about AI deception, I explain why misaligned behaviors emerge from reinforcement learning,…
In an interview in this article for The @guardian about AI deception, I explain why misaligned behaviors emerge from reinforcement learning, why they will continue to pose risks as models become more
In an interview in this article for The @guardian about AI deception, I explain why misaligned behaviors emerge from reinforcement learning, why they will continue to pose risks as models become more capable, and how we intend to rethink how we train AI systems at @LawZero_ . https://www.theguardian.com/news/2026/sep/01/if-you-build-something-vastly-smarter-than-you-it-better-be-on-your-side-can-we-stop-ai-from-deceiving-us
Related
- Another real-world manifestation of the kind of misaligned actions frontier systems developed by leading companies can take to achieve goals…
- I was happy to join the @ScienceBoard_UN podcast to discuss deception among frontier AI models and the need to manage its potential global i…
- Reinforcement Learning Amplifies Emergent Misalignment from Harmless Rewards
Source: Yoshua Bengio (X) | 2026-09-01