An escalating conflict between Washington and the artificial intelligence industry reached a breaking point Friday. Anthropic abruptly disabled all global access to its newly released Fable 5 and Mythos 5 AI models to comply with a sweeping export control directive from the Trump administration. The sudden restriction targets foreign nationals, cutting off hundreds of millions of users worldwide from the frontier technology.
The government ordered the shutdown over a specific security vulnerability. Authorities cited a narrow potential jailbreak they feared could aid computer hackers. Anthropic received the government letter at 5:21 p.m. ET on Friday, a timeline verified by a detailed Washington Post report. Rather than attempting to screen the nationality of individual global users, Anthropic took both models completely offline.
Key Details
- The export restriction bans any foreign national from accessing the models, including Anthropic’s own legally employed staff working inside the United States.
- Fable 5 launched publicly just days prior as a heavily guardrailed model.
- Mythos 5, an advanced cybersecurity system, was already tightly restricted to vetted users like U.S. cyber defenders prior to the ban.
- All older Anthropic models remain online and unaffected by the directive.
The abrupt export control follows recent executive action from the White House. President Donald Trump signed an executive order just 10 days earlier establishing a voluntary framework for the federal government to vet national security risks of advanced AI systems for up to a month before public release. The mandatory shutdown of Fable 5 bypasses that voluntary structure.
Anthropic pushed back against the severity of the government’s technical claims. In its defense, the company argued that the narrow jailbreak cited in the directive produced only minor, previously known vulnerabilities. Anthropic explicitly noted that these same vulnerabilities are easily surfaced by publicly available competitor models, specifically naming OpenAI’s GPT-5.5.
How the Export Ban Escalates the Pentagon’s AI Supply Chain Fight
The Friday directive stands as the U.S. government’s most aggressive step to date to impose direct export controls on frontier AI models. The restriction enforces a strict personnel and usage rule change across the tech sector. By explicitly banning foreign nationals from using specified models, the rule upends Anthropic’s internal engineering operations and global API deployments.
This shutdown directly escalates an ongoing legal and operational fight between Anthropic and federal agencies. Earlier in 2026, the Department of Defense declared the company a supply chain risk, requiring defense contractors to certify they would not use Claude in military applications. Anthropic sued the administration over that initial designation. The new Trump administration directive bypasses procurement rules entirely to directly restrict commercial access to the company’s newest technology.