Policymakers grapple with fallout from Chinese AI-enabled hack

Policymakers and security firms are responding to reports that Chinese actors jailbroke Anthropic’s Claude model and used it to automate a cyberespionage campaign against more than 30 entities, prompting concerns about AI accelerating attacks, calls for stronger model safeguards and defensive AI, and debate among experts who note both the potency and current limitations of AI-assisted hacking.