The U.S. government has implemented a suspension of foreign access to Anthropic's AI models, specifically the Fable 5 and Mythos 5, following concerns raised by Amazon CEO Andy Jassy. This decision, announced on Friday, was influenced by Jassy's communication with senior officials regarding potential security risks identified by Amazon researchers.
On Thursday, researchers found a method to exploit the Fable 5 model, prompting Jassy to alert the administration. The subsequent response involved a rapid evaluation of the threat and outreach to Dario Amodei, CEO of Anthropic, who reportedly resisted the administration's requests for voluntary compliance. David Sacks, co-chair of the President’s Council of Advisors on Science and Technology, indicated that the administration acted reluctantly, expressing surprise at Anthropic's lack of cooperation in addressing the identified vulnerabilities.
Anthropic, which claims around 18,900 monthly users for its Claude AI, stated that it believes the U.S. directive stemmed from a misunderstanding. The company is currently working to resolve the issues and restore access to its models for users.