Emergency Suspension of AI Capabilities
Anthropic, a leading artificial intelligence company, announced on June 13, 2026, that it has been forced to abruptly disable its most advanced AI models, Fable 5 and Mythos 5. The decision follows an urgent export control directive issued by the US Department of Commerce, which prohibits access to these models for foreign nationals due to unspecified national security concerns.
The company stated that it received the order on Friday at 5:21 p.m. ET. According to Anthropic, the government’s intervention is based on a belief that a method exists to “jailbreak” the models—specifically, bypassing safety guardrails to identify software vulnerabilities. While the US government has only provided verbal evidence of a narrow potential exploit, the directive mandates a total suspension of these models to ensure compliance.
Tensions Over AI Security and Regulation
This development marks a significant escalation in the ongoing friction between the Trump administration and Anthropic. Earlier this year, the US government designated Anthropic a “supply-chain risk,” a move the company is currently challenging in court. The designation followed the company’s refusal to integrate its AI models into domestic surveillance and autonomous weapons systems.
Kirsten Davies, the Pentagon’s chief information officer, confirmed the government’s stance on social media, stating: “Some things are simply more important than revenue cycles, clickbait and pre-IPO valuation. America First. Always.”
Analysis: The Shift in Regulatory Strategy
The directive represents a pivot in US strategy. Historically, export controls focused on the physical infrastructure of AI—chips and hardware. By restricting access to the software models themselves, the government is signaling that the intelligence capability of the models is now treated as a strategic asset. Critics, however, argue that this “recall” approach is disproportionate. Anthropic has noted that the vulnerabilities identified are “relatively simple” and that similar capabilities exist in competing models from other providers.
The economic stakes are high: Anthropic recently filed for an IPO, and the loss of its flagship “Mythos-class” models could impact its valuation and market standing. As the industry grapples with these new boundaries, the case highlights the lack of a standardized framework for defining “unacceptable risk” in AI deployments, potentially creating a chilling effect on international research collaboration.