Anthropic’s safety warnings may have just backfired — the government has pulled the plug on its most powerful AI

Anthropic has expressed public disagreement with a government decision to restrict access to one of its advanced artificial intelligence models following the discovery of a potential security vulnerability. The company argues that the narrow nature of the identified jailbreak does not justify a product recall, while regulators maintain that safety protocols must prioritize preventing unauthorized system access. This dispute highlights an ongoing tension between AI developers and government agencies regarding the thresholds for acceptable risk in public-facing large language models.