US Lifts Anthropic AI Export Controls

The US Department of Commerce lifted emergency export controls on Anthropic's Claude Fable 5 and Claude Mythos 5 AI models, which had been imposed on June 12 due to national security concerns. The controls were enacted after Amazon researchers reported a method to bypass Fable 5's cybersecurity safeguards, allowing it to identify software vulnerabilities and generate exploit code. Anthropic, unable to verify user nationality in real time, suspended global access to both models. Following collaboration with the United States — United States Department of Commerce and the Artificial intelligence safety institute (CAISI), Anthropic introduced an AI-powered safety classifier that blocks the reported jailbreak technique in over 99% of cases, redirecting blocked requests to an older model. Global access to Claude Fable 5 was restored on July 1, while Claude Mythos 5 remains under tighter controls, available only to vetted US organizations through Project Glasswing. Anthropic also announced new safety measures, including a HackerOne bug bounty program, expanded pre-release testing, and closer cooperation with the US government, and is working with Amazon, Microsoft, and Google to develop an industry framework for assessing AI jailbreak severity. The company is also implementing Know Your Customer (KYC) procedures with Persona for certain Claude services.