Amazon CEO Andy Jassy and other tech leaders warned senior Trump administration officials about security risks in Anthropic's most advanced AI models this week, a source familiar with the matter said. The warnings preceded a rare federal export control order that forced Anthropic to globally disable its latest models after the company refused to fix a reported "jailbreak" vulnerability.
The security dispute
Anthropic had previously warned about the hacking capabilities of its Mythos model and held it back from wide release. Earlier this week, the startup rolled out a public version called Fable, describing it as equipped with cybersecurity safeguards. That brief release ended Friday when the U.S. government informed Anthropic it believed a method existed to bypass those safeguards. According to Anthropic, the bypass only exposed "minor" security flaws that other publicly available models can also find.
Federal intervention and export controls
The administration ordered Anthropic to block foreign nationals, both inside and outside the U.S., from using Fable 5 and Mythos 5. In response, the company disabled global access to both models. White House adviser David Sacks said officials issued the export control reluctantly after Anthropic CEO Dario Amodei "refused" to "fix the jail break or de-deploy the model." Sacks added that the administration hopes Anthropic remediates the safety issue so the export control can be lifted and Fable can return to general release.
Amazon did not confirm its specific discussions with government officials. An Amazon spokesperson said the company does not share details of such conversations, noting that governments frequently seek counsel from leading cloud providers on potential security risks.
Industry and expert reaction
The broad nature of the restriction has drawn criticism from experts who support targeted export controls on advanced AI. "This was not well thought-out," said Jimmy Goodrich, a senior fellow at the University of California's Institute for Global Conflict and Cooperation. "It even bans Canadians and Brits employed at Anthropic from doing research and development." The Information, a technology news outlet, cited a U.S. official reporting that the administration is unlikely to force other AI firms to abide by restrictions similar to those placed on Anthropic.
For officials addressing these jurisdictional overlaps, frameworks for AI for Policy Makers highlight the challenges of balancing national security with international research collaboration. As federal agencies calibrate their approach to emerging technologies, tracking AI for Government initiatives will be critical for understanding how export controls influence domestic innovation.
Why this matters for government professionals
Federal export controls on AI models directly affect how agencies evaluate vendor risk and manage international research partnerships. This incident demonstrates that national security mandates can override corporate release schedules, requiring procurement and policy teams to anticipate sudden shifts in software availability. Officials must build contingency plans for critical workflows that rely on third-party AI tools, ensuring compliance with shifting federal directives.
Your membership also unlocks:
