The Korean Telecom Giant at the Center of Anthropic's Mythos Controversy (www.wired.com)

Anthropic's access to its Claude Mythos AI model has been revoked by the Trump administration after concerns emerged regarding South Korean telecom giant SK Telecom's historical ties to China. This decision followed a notification from Amazon to the White House about vulnerabilities in Anthropic's Fable 5 model, which could potentially allow unauthorized access to powerful cyber capabilities. Despite Anthropic's arguments that these vulnerabilities are not unique to its models, the administration concluded that it could not trust Anthropic to protect its advanced technology adequately, forcing the company to disable Mythos and Fable 5 access for all foreign nationals. This situation is significant for the AI/ML community as it underscores the growing intersection between national security and AI technology management. Anthropic, which had been expanding access to its models under Project Glasswing, found itself having to restrict access entirely rather than implement complex nationality-based access controls. The situation reflects the rising scrutiny of company affiliations with foreign entities, particularly amid concerns over cybersecurity and geopolitical tensions. The incident is a stark reminder of the implications such collaborations can have on the advancement and accessibility of cutting-edge AI technologies in an era marked by cautious regulatory oversight.