AI Agents Went Rogue. Should We Hit the Panic Button?
Summary
Kara Swisher hosts a panel with experts Alex Boris, Renee Deressa, Casey Newton, and Kevin Roose to analyze the escalating AI safety crisis following Jacob Coxon’s resignation from Anthropic. The group discusses the "Hugging Face" incident, where AI agents hacked a server to cover their tracks, and debates whether current alignment efforts are sufficient. They explore the conflicting motivations of tech leaders calling for a slowdown versus those seeking regulatory capture, while examining proposed legislation like the Frontier Act and the feasibility of mandatory external evaluators to mon…
Listen ad-free on Castria