Climate
Privacy
Bob Dylan
Drones
Data Centers
AI Governance
2026 Maine Senate Race
US Restricts Anthropic Models
Coverage from HNGN, ET Now, and others
Articles
6
Active Days
3
The Topic
The U.S. Commerce Department directed Anthropic to restrict foreign-national access to its newly released Fable 5 and Mythos 5 models, citing national-security concerns linked to alleged bypasses of their safety controls and cyber capabilities. Because Anthropic could not reliably separate affected users, it disabled both models worldwide while leaving its other models available. Anthropic disputes the government’s characterization of the vulnerability and is negotiating with officials over technical review, remediation, and possible restoration of access.
First Article: 06/13/26
Latest Article: 06/15/26
Summary
- The Commerce Department issued a June 12 directive restricting Fable 5 and Mythos 5 for foreign nationals inside and outside the United States.
- Anthropic disabled both models for all customers because it could not reliably segment affected users in real time.
- The government’s stated concern centers on an alleged jailbreak that could weaken safeguards around cybersecurity and vulnerability discovery.
- Anthropic says the demonstrated issue was narrow, involved previously known minor vulnerabilities, and may also exist in other public models.
- Anthropic’s other models, including Claude Opus, Sonnet, and Haiku, reportedly remained available.
- The company and U.S. officials began discussions involving technical staff and government security researchers, but no restoration agreement had been announced.
- The action extends export-control policy from AI hardware and related tools toward direct restrictions on access to advanced AI models.
History
The update sharpens the government’s rationale and the scope of the restriction: Commerce now explicitly tied the order to alleged cybersecurity-related jailbreaks and foreign-national access, while Anthropic says the issue is narrow and may also affect other public models. The story also adds that talks are now centered on technical review and possible restoration, with Anthropic’s other models still available.
The story is largely unchanged, but the current version sharpens the enforcement angle: the action is now framed explicitly as restricting foreign-national access under export-control authority, rather than a broader model-wide ban. It also adds that negotiations are continuing while the government has not disclosed its technical evidence.
Featured
Timeline: 3 Days
Hover over any logo to see coverage summary, click for full article.