Anthropic has been forced by the US government to suspend all foreign access to its latest AI models, Claude Fable 5 and Mythos 5, marking a significant escalation in tensions between the AI company and federal regulators regarding national security concerns.
- Anthropic disables Fable 5 and Mythos 5 after US export control order
- Government cites national security but provides no detailed vulnerabilities
- Tensions escalate following US designation of Anthropic as supply-chain risk
What happened
Just days after Anthropic released its most advanced AI models, Claude Fable 5 and Mythos 5, the company received an export control directive from the Trump administration compelling it to suspend all access to these models for any foreign nationals, whether located inside or outside the US. Anthropic promptly disabled the models for all users to comply with the government mandate.
This directive came as an acceleration of ongoing scrutiny by the US government, which had earlier classified Anthropic as a supply-chain risk to national security. The government referenced concerns related to national security in the order but did not provide specifics. Anthropic noted that the directive followed a demonstration of a jailbreak technique used to expose minor vulnerabilities in the Fable 5 model.
Why it matters
The US government’s intervention highlights escalating national security concerns regarding the proliferation and use of powerful AI technologies across borders. By targeting Anthropic’s latest models, the administration is signaling tighter controls on AI exports amid fears about their potential misuse or exploitation by foreign entities.
Anthropic’s position is precarious as it balances government contracts and requests for safety assurances with compliance to export restrictions. The dispute has broader implications for AI innovation, international collaboration, and the regulatory landscape shaping the development and deployment of cutting-edge AI systems.
What to watch next
Stakeholders should monitor if the US government expands export controls to other AI companies or model versions, and whether there will be clarifications or legal challenges regarding the vagueness of the national security rationale. The response from other AI firms and international users affected by such measures will also be critical.
Anthropic’s future releases and ongoing negotiations with the government regarding export regulations and compliance frameworks will be important to track, as they may set precedents for AI safety governance and the balance between innovation and security in the rapidly evolving AI sector.