Anthropic Disables Fable 5 AI Model Worldwide After U.S. Export Order

U.S. Government Forces Global Shutdown of Advanced AI Models

Anthropic abruptly disabled access to its latest AI models, Fable 5 and Mythos 5, for all customers worldwide after receiving a U.S. government order restricting access to non-American users. The company announced the suspension Friday evening, stating it had no choice but to comply immediately.

"The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance," Anthropic said in its statement.

The move affects both domestic and international users, despite the export controls initially targeting only non-U.S. access. Fable 5, which launched just this week, is a less powerful variant of Anthropic's Mythos model designed with additional safeguards for general commercial use.

Why It Matters

The shutdown demonstrates how U.S. export controls on advanced AI can create immediate global disruptions, even for American companies' domestic customers. The incident has intensified calls for technological sovereignty in Europe and the U.K., where governments now face sudden loss of access to cutting-edge AI capabilities they may have integrated into operations. It also highlights the growing tension between AI safety concerns and commercial deployment as models become more capable.

Jailbreaking Concerns Drive Decision

According to Anthropic, the U.S. government imposed the export controls over concerns that the models' built-in safety mechanisms could be circumvented through jailbreaking—techniques that bypass AI guardrails to enable prohibited uses.

However, Anthropic pushed back on the rationale, stating the government had only provided "verbal evidence of a potential narrow, non-universal jailbreak." The company argued that discovering a limited jailbreak vulnerability should not warrant recalling an entire commercial model.

Mythos, the more powerful model underlying Fable 5, was initially withheld from broader release due to Anthropic's own concerns about its ability to exploit software vulnerabilities. In early April, the company granted limited access to Mythos Preview through its Glasswing project, restricting it to trusted technology and cybersecurity firms to help strengthen defensive capabilities.

International Fallout and Sovereignty Debates

The suspension left non-U.S. governments and institutions, including EU bodies, suddenly cut off from access they had been seeking. The decision has fueled discussions about reducing dependence on American AI infrastructure.

"As we debate the future of national security and technological sovereignty, access to AI capabilities is crucial," said Kanishka Narayan, the U.K.'s minister for AI, responding to the development.

Fable 5 was marketed as a "Mythos-class model" with enhanced safeguards to prevent misuse of its advanced capabilities, particularly in cybersecurity applications. At launch, Anthropic warned that "without safeguards, Fable 5's capabilities in areas like cybersecurity could be misused to cause serious damage."

These details were first reported by Politico.