NewsPublished 2 Jul 2026

Trump Administration Lifts Restrictions on Anthropic’s Claude Models After Cybersecurity Alarm

Source SecurityWeek

Description

Anthropic said Tuesday night that its AI model called Claude Fable 5 is now widely available. The post Trump Administration Lifts Restrictions on Anthropic’s Claude Models After Cybersecurity Alarm appeared first on SecurityWeek .

Impact

The AI models demonstrated the capability to identify software vulnerabilities, which could potentially be weaponized by malicious actors to compromise critical computer networks. This raised national security concerns leading to temporary restrictions and regulatory oversight. However, there are no reports of active exploitation or known attacks in the wild related to these vulnerabilities. The impact is primarily related to the potential misuse of AI-powered vulnerability discovery rather than a direct software flaw or exploit.

Defensive guidance

The U.S. government imposed temporary restrictions on access to Anthropic's AI models to mitigate national security risks. Anthropic complied by taking the models offline for all users except approved U.S.-based organizations. The restrictions on Claude Fable 5 have now been lifted, and the model is widely available again, indicating that the immediate cybersecurity concerns have been addressed or sufficiently mitigated. Access to Mythos 5 remains controlled. Organizations should follow government guidance and only use these AI models if authorized. No additional mitigation actions are specified or required at this time.

Summary

Anthropic's AI models Claude Fable 5 and Mythos 5 were temporarily restricted by the U.S. Commerce Department due to cybersecurity concerns after Amazon researchers found a way to bypass safeguards in Fable 5, enabling discovery of software vulnerabilities that could be weaponized. The Trump administration's executive order on AI oversight established a framework for vetting advanced AI systems for national security risks before public release. Following this review, restrictions on Claude Fable 5 were lifted, making it widely available, while Mythos 5 remains accessible only to approved U.S. organizations. The models' ability to find software flaws raised concerns about potential exploitation of critical computer networks. OpenAI also limited access to its new AI model during this period. No confirmed exploits or attacks leveraging these vulnerabilities have been reported.

References

securityweek.com/trump-administration-lifts-restrictions-on-anthropics-claude-models-after-cybersecurity-alarm/ (opens in a new tab)

Community reviews

Loading reviews…

Description

AI analysisMachine-generated, updated 2 Jul 2026, 11:07 UTC

References

Community reviews