Anthropic has suspended access to two of its most advanced AI models after the U.S. government ordered the company to prevent foreign nationals from using them over national security concerns.
The directive applies to Claude Fable 5 and Claude Mythos 5, including access by foreign nationals working at Anthropic. The company said it had to disable both models for all customers to comply, while access to its other models remains unchanged.
Government Raises Cybersecurity Concerns
Anthropic disclosed the order in an official statement published shortly after Fable 5 became publicly available. The company said authorities had not provided a detailed explanation of the security concerns behind the directive.
Anthropic believes the decision relates to a method for bypassing Fable 5’s safeguards, commonly known as jailbreaking. Such techniques use prompts or other methods to get around restrictions placed on an AI system.
The company said it reviewed a demonstration in which the technique helped identify a small number of previously known and minor software vulnerabilities. Anthropic added that other publicly available AI models could find the same weaknesses without using the bypass.
Fable 5 Had Limited Early Access
Before its wider release, Fable 5 was privately available to a small group of organizations for testing. Anthropic had warned that the model’s cybersecurity abilities were stronger than those of any model it had previously made generally available.
The company introduced additional safeguards because the model could identify or exploit weaknesses in computer systems. Some critics questioned whether Anthropic’s description of the model as too powerful for an unrestricted release was partly promotional.
Finance, technology, and government officials had also raised concerns before the public launch. The U.K. AI Security Institute reportedly found during testing that the model could exploit cybersecurity defences and systems in 73% of attempts.
Separate Pentagon Dispute Continues
The suspension comes as Anthropic remains involved in a separate legal dispute with the Trump administration. The Pentagon previously designated the company a supply chain risk after disagreements over limits on military uses of Claude.
Anthropic challenged the designation, arguing that its requested restrictions concerned mass domestic surveillance and fully autonomous weapons. A federal judge temporarily blocked enforcement of the Pentagon directive while the lawsuit continues.
The new export control order is separate from that case. TechCrunch reported that the government’s action required Anthropic to withdraw access only days after Fable 5’s release.
European Officials Assess The Decision
The European Union gained access to Mythos earlier in June after several weeks of discussions. European Commission spokesperson Thomas Regnier said officials were assessing Anthropic’s announcement.
Regnier said the development highlighted Europe’s need for greater technological sovereignty. The commission recently announced measures intended to reduce the region’s dependence on U.S. and Asian suppliers for technologies including AI.
Gina Neff, professor of responsible AI at Queen Mary University of London, said restricting access could limit testing and international collaboration. She added that governments and companies need stronger capabilities to protect their systems as AI cybersecurity tools improve.
Featured image credits: Heute.at
For more stories like it, click the +Follow button at the top of this page to follow us.