Anthropic, a prominent player in the field of artificial intelligence, announced on Friday that it has temporarily disabled its latest AI models, Fable 5 and Mythos 5, in response to a directive from the Trump administration aimed at preventing their usage by individuals from foreign countries.
The imposition of export controls represents a significant move by the U.S. government to limit access to cutting-edge AI models. While Anthropic made Fable publicly available earlier this week, Mythos, a more advanced version, remains restricted due to concerns over cybersecurity.
In a released statement, Anthropic expressed its disagreement with the government’s decision, noting that the directive received on Friday did not specify the exact national security apprehensions prompting the action.
The company argued that a transparent and fair process, based on technical facts, should be in place for blocking unsafe deployments rather than unilateral directives. Anthropic labeled the situation as a “misunderstanding” and expressed a desire to reinstate access to the models promptly.
The Commerce Department did not immediately respond to a request for comment.
This move comes shortly after President Donald Trump signed an executive order mandating a month-long evaluation by the federal government of the national security implications of advanced AI systems before their public release. Participation in this evaluation process is voluntary for AI developers.
During the unveiling of Mythos in April, Anthropic highlighted the model’s capacity to identify software vulnerabilities at scale, underscoring the potential risks to economies, public safety, and national security if the technology is misused.
-
Cybersecurity experts and the banking sector express concerns over Anthropic’s latest AI model. Learn more.
-
AI leader Anthropic confidentially files for IPO in the U.S.
Bank of Canada governor Tiff Macklem noted discussions about Mythos during the International Monetary Fund’s spring meetings, emphasizing the ongoing need for proactive planning to address the evolving landscape of AI technology.
Macklem stressed the importance of robust cybersecurity measures in light of the potential of AI models like Mythos to expose and exploit vulnerabilities, highlighting the necessity for preparedness among firms, regulators, and policymakers.