American technology company Anthropic has temporarily disabled its advanced AI models, Claude Fable 5 and Claude Mythos 5, following export control directives from the U.S. government. This unexpected move highlights the increasing impact of national security considerations on the tech industry. The new directive restricts access to specific AI technologies for all foreign nationals, regardless of their location or position, even if they work for the company domestically.
The U.S. government cited national security concerns as the primary reason behind the directive but did not provide specific details. Instead, the concern seems to stem from potential vulnerabilities or "jailbreaking" methods that may allow unauthorized access or modification of these models. Despite these concerns being described as minor and relatively easily detectable by other models, the government has insisted on restrictive measures.
Anthropic, however, expressed disagreement with Washington’s decision, arguing that the reasons provided do not justify the commercial halt of Claude Fable 5 and Mythos 5. The company acknowledges that the identified vulnerabilities were straightforward and not unique to their models. They even mentioned that the vulnerabilities could be identified without employing any advanced techniques.
Only recently launched in June 2026, Claude Fable 5 has already established itself as Anthropic’s most powerful model, excelling in programming, data processing, document analysis, and scientific research. Despite its capabilities, specific requests within cybersecurity, biology, chemistry, and model distillation are redirected to another less powerful model, Claude Opus 4.8, upon detection of potential threats.
Meanwhile, Claude Mythos 5 is reputed as one of the most advanced models in programming and agent-based artificial intelligence. The model is designed to self-execute complex tasks and identify cybersecurity vulnerabilities, a feature that has attracted interest from the U.S. National Security Agency despite the Defense Department flagging Anthropic as a risky supplier.
As the government continues to navigate the intricate balance between technological innovation and national security, this development raises important questions about the future of AI technology and its regulation, both in the U.S. and globally.
