Back

US Tightens AI Control: Anthropic Faces Restrictions on Advanced Models Under New Export Rules

Brief By Newsbrief / 12:49 PM on 14 Jun 2026


The United States has taken a major regulatory step in artificial intelligence governance by issuing new export control restrictions affecting advanced systems developed by Anthropic. Under the new rules, access to certain high-end AI models has been restricted for foreign nationals and non-US entities, marking one of the strictest measures introduced so far in the AI sector.

According to the updated directive, access to advanced models such as Fable 5 and Mythos 5 has been suspended for all foreign users, including employees of Anthropic who are not US citizens. The restriction applies globally and has been implemented under national security considerations cited by US authorities. As a result, the company has been instructed to immediately halt services for these models to ensure compliance with export control regulations.

Anthropic confirmed that the directive directly impacts the availability of Fable 5 and Mythos 5, both of which were positioned as advanced AI systems designed for specialized applications. The company stated that Mythos-level AI technology was developed with capabilities aimed at government and enterprise use cases, including cybersecurity analysis and complex software code detection. Fable 5 was also described as part of this advanced AI lineup intended for controlled deployment environments.

However, Anthropic clarified that other Claude AI models remain unaffected by the new restrictions and continue to be available for users. The company has expressed regret over the disruption caused and suggested that the decision may stem from a misunderstanding of operational frameworks. It further indicated that discussions are underway to resolve compliance issues and restore access to the restricted services.

The move highlights growing global concerns around AI governance, security, and cross-border technology access, as governments increasingly seek tighter control over advanced artificial intelligence systems.

Read more on Newsbrief