Anthropic, the AI company behind Claude, was forced to disable its most advanced AI models—Claude Fable 5 and Mythos 5—just days after their launch, following a US government export control directive. The order, issued by the Department of Commerce, bars all foreign nationals, including Anthropic's own foreign employees, from accessing the models, citing national security. Anthropic complied by shutting down both models globally, as it could not filter users by nationality. The company expressed frustration, stating the government provided no specifics and that the alleged security vulnerability (a jailbreak technique) was minor and also present in other AI models. The incident escalates a long-running conflict between Anthropic and the Trump administration, which earlier this year designated the company a 'supply chain risk' and limited its use by federal agencies. Anthropic has sued the administration over that designation. The suspension has drawn international attention, particularly in France, where officials voiced alarm over extraterritorial implications.
Lack of transparency: Spanish outlet criticizes the government's vague national security justification.
El Diario focuses on Anthropic's complaint that the government did not provide specific national security concerns and that the action lacks transparency. The article emphasizes that Fable 5 had safety guardrails and that Mythos 5 was restricted to vetted partners via Project Glasswing.
Broad restrictions affecting allies: Argentine newspaper highlights the wide scope including employees from allied nations.
ClarÃn reports on the order's unusually broad restrictions, noting it could affect employees from Canada, UK, etc. The article recounts the earlier Pentagon designation and the concern that Mythos could be a powerful hacking tool, though it was shared with 40 organizations for defensive purposes.
International alarm: French magazine raises questions about extraterritorial implications for allied nations.
L'Obs presents the story through a Q&A format, detailing the models and the order's unprecedented scope. It cites a French political reaction calling the directive alarming, and notes the order came from Commerce Secretary Lutnick based on a jailbreak discovery.
Technical dispute: In-depth look at the jailbreak claim and Anthropic's rebuttal.
Wired provides detailed coverage of the government's alleged jailbreak technique, quoting Anthropic's blog post that the vulnerability is minor and not unique. It also describes the security features of Fable 5 and the limited rollout of Mythos. The article emphasizes the unprecedented nature of the order.
Timeline and impact on users: Anthropic pulls models abruptly due to last-minute government order.
Mashable reports on the sudden removal of Fable 5 and Mythos 5, highlighting the timing of the order (5:21pm ET) and Anthropic's compliance. It notes the previous supply chain risk designation and Anthropic's view that the jailbreak vulnerability is simple and shared by other models.
Business and IPO impact: Italian business paper links the order to Anthropic's upcoming IPO and financial implications.
Il Sole 24 Ore highlights the business consequences, noting Anthropic recently filed for IPO. It details the order's timing and the company's argument that the government action violates fair regulatory principles. The article also references the Pentagon's supply chain risk designation.
The US government's unprecedented order to block foreign access to Anthropic's AI models marks a significant escalation in AI regulation, blending export controls with national security. While the administration frames this as a necessary safeguard, Anthropic and critics argue the process lacks transparency and factual basis, potentially harming US competitiveness and global trust. The incident underscores the growing tension between AI companies seeking responsible deployment and a government increasingly assertive about controlling advanced technology. The outcome may influence how other AI firms handle sensitive models and could affect Anthropic's upcoming IPO.
Specific origin of the government order
| Outlet | Claim |
|---|---|
| L'Obs | The directive came from Commerce Secretary Howard Lutnick, acting after learning a company had bypassed safety guardrails. |
| Mashable | The order is from the Trump administration via export control directive, no specific official named. |
| Il Sole 24 Ore | Washington ordered the suspension under export control regulations; no individual named. |
The reporting reveals a sharp divide between the government's apparent concern about advanced AI capabilities falling into foreign hands and Anthropic's insistence that the process was arbitrary and the technical threat overstated. While the order is an extraordinary use of export controls targeting a domestic AI company's entire user base, the lack of public evidence from the government weakens its case. The incident may set a precedent for more aggressive AI regulation, but it also risks alienating allies and stifling innovation if transparency is not improved. The alignment of most outlets in questioning the order's proportionality suggests skepticism among media, though none outright defend foreign access.
Get tomorrow's top stories in your inbox