Anthropic blocks own models after Trump export order
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic blocked access to its Fable 5 and Mythos 5 models for all foreign nationals, including US-based users and its own employees, after the Trump administration issued an export control directive citing "national security authorities."
- The Trump administration has not publicly explained the legal basis for the order; Anthropic said the government raised concerns about a potential "jailbreak" by China-linked groups, but the company claimed that jailbreak did not let users bypass all of its safeguards.
- Hanna Dohmen, a senior research analyst at Georgetown's Center for Security and Emerging Technology, called it "an open question" whether the order strains existing rules, and said it appears to be "the first time US export controls have been used to control access to an AI model in this way."
- Fable and Mythos remain hosted on Anthropic's servers and users receive only chatbot responses — not source code or model weights — making the "export" framework unclear, since export controls traditionally apply to discrete items shipped across borders rather than remote service access.
- Andrew Reddie, a professor at UC Berkeley's Goldman School of Public Policy, said the episode "makes clear the unsustainability of the existing governance regime" and warned that requiring jailbreak-proof models as a de facto standard would leave the United States with "no AI models."
- The order raises parallel questions for other frontier labs including OpenAI, Google, Meta, and xAI, since experts say it remains unclear whether the action targets specific model capabilities, safeguard issues, or Anthropic's reportedly strained relationship with the Trump administration.
Why it matters: If the Trump administration can unilaterally force Anthropic to yank its frontier models without explaining the legal basis, every major AI lab faces the same opaque, ad hoc risk on every new release. The episode contradicts the administration's stated hands-off AI stance and has fueled arguments that foreign governments and companies should avoid relying on American AI providers for strategically important systems.
