The U.S. government has ordered Anthropic to suspend access to its newly released Claude Fable 5 and Mythos 5 models over national security concerns regarding a safety jailbreak technique.
Anthropic has suspended access to its recently launched Claude Fable 5 and Mythos 5 models in compliance with a U.S. government directive. The order was issued after the government became aware of a bypass technique targeting the models' safety guardrails. Anthropic has complied with the restriction while noting that its own review found the bypass only exposes minor, previously known vulnerabilities similar to those present in other public models, rather than a novel or universal jailbreak.
This unprecedented government-mandated shutdown of a commercial frontier AI model marks a major escalation in state intervention under the guise of national security.
* The action demonstrates the government's willingness to invoke export control and national security powers to halt AI access over relatively minor or known vulnerabilities.
* The suspension highlights the tension between AI labs' internal safety assessments and the government's risk tolerance.
* This move will likely accelerate debates around the transparency of government AI directives and the criteria used to determine national security threats.
DISCOVERED
1h ago
2026-06-13
PUBLISHED
2h ago
2026-06-13
RELEVANCE
AUTHOR
bridgemindai