White House forces Anthropic’s Fable AI models offline over security fears

A frantic day of pressure inside the White House ended with Anthropic taking its newest flagship models offline worldwide, turning a technical safety fight into a test of presidential power over AI. Senior officials moved after concerns that Fable 5 and Mythos 5 could be accessed in ways that widened national security risks, including fears tied to China-linked technology.

Anthropic said it received the government directive on June 12, 2026, at 5:21 p.m. ET. The order required suspension of access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. Anthropic said it disabled both models for all customers around the world to comply, while its other models remained available.

The clash had been building for months. On February 27, 2026, President Trump directed federal agencies to stop using Anthropic technology, and Defense Secretary Pete Hegseth announced a supply-chain risk designation after a prolonged dispute between the Pentagon and Anthropic over how the company’s AI systems could be used. Anthropic had pushed back against Pentagon requests for uses in fully autonomous weapons and mass domestic surveillance, arguing that frontier models were not reliable enough for autonomous weapons and that mass surveillance of Americans would violate fundamental rights.

Even as that relationship soured, there were signs of a possible thaw in late April, when the White House was reportedly drafting guidance that could allow agencies to work around the supply-chain risk designation and resume using Anthropic tools. But the release of Fable 5 and Mythos 5, launched only days before the shutdown, quickly became the center of another confrontation. Anthropic said it had spent thousands of hours red-teaming the models with the U.S. government, the UK AI Security Institute, private third-party organizations and its own teams before launch, and said no tester had found a universal jailbreak.

Instead, the company said the government never gave specific details of its national-security concern, though Anthropic believed officials had become aware of a possible jailbreak method for Fable 5. Anthropic said it reviewed the technique and found only a small number of previously known, minor vulnerabilities. The broader concern, according to reporting on the dispute, was that a China-linked group may have accessed Mythos 5, adding a geopolitics layer to a conflict Anthropic had framed as a safety issue.

The episode now stands as a warning to the AI industry: once model access is tied to export controls and national security, a product can be forced offline as fast as it is launched.