White House Imposes Export Controls on Anthropic's Mythos Model Over China Security Concerns

Summary

The Trump administration imposed export controls on Anthropic's advanced Mythos AI model and its consumer version Fable 5, restricting access to US citizens only. According to sources, the decision was partly driven by suspicions that a China-linked group had accessed Mythos, raising national security concerns about potential reverse-engineering and AI capability theft through a process known as model distillation. The restrictions were so stringent that they would have prohibited Anthropic's own foreign national employees from using the models.

The White House also raised concerns about a jailbreak vulnerability in Fable 5 that could bypass its safety guardrails. Trump adviser David Sacks alleged that when the administration notified Anthropic of the vulnerability, CEO Dario Amodei dismissed it as not a serious risk and refused to fix it. Amazon's CEO Andy Jassy reportedly alerted the government to the security risks, highlighting concerns that the model could be exploited by malicious actors.

Rather than comply with the export restrictions, Anthropic chose to remove both Mythos and Fable 5 from the market entirely. The decision marks an escalation in tensions between Anthropic and the Trump administration over AI safety standards and government regulation. Anthropic has previously clashed with the administration over autonomous weapons policy and state AI regulation.

• Anthropic's decision to withdraw models entirely rather than restrict access reflects escalating tension with Trump administration over AI regulation and safety priorities