The Trump administration ordered Anthropic to revoke SK Telecom's access to Claude Mythos over alleged China ties. Anthropic chose to disable the models entirely rather than implement nationality-based access restrictions.
The Trump administration ordered Anthropic to revoke access to its most advanced AI models, Claude Mythos and its public version Fable 5, for all foreign nationals, including immigrants in the US. Anthropic decided to disable the models entirely rather than implement nationality-based restrictions.
The US government raised concerns about SK Telecom's alleged ties to China. SK Telecom was one of about 150 organizations granted early access to Mythos. The situation escalated when Amazon researchers reported vulnerabilities in Fable 5 that could bypass guardrails and access Mythos's cyber capabilities.
This incident highlights the tension between advanced AI export controls and national security. Anthropic remains in negotiations with the White House but has not yet reached an agreement to restore the models. It underscores the challenges AI companies face navigating government regulations and global partnerships.
Comments focus on China's AI model progress and distillation controversies, as well as Anthropic's decision to hide reasoning tokens. Some predict China will soon release comparable models, citing restrictions and hardware limitations as obstacles. Critics also argue that Wired's article exaggerates the SK Telecom issue, suggesting the real reason for the shutdown was Fable's jailbreak vulnerabilities.