Amazon expresses concerns about artificial intelligence models from Anthropic

An informed source told Reuters that Amazon CEO Andy Jassy was among the technology sector leaders who, over the past few days, expressed their concerns to senior officials in President Donald Trump’s administration about security risks in Anthropic’s most advanced artificial intelligence models.

Jassy’s participation highlights the exceptional step Anthropic took on Friday to discontinue its latest models globally in response to national security orders issued by the Trump administration.

The emerging artificial intelligence company, based in San Francisco, had previously warned of the hacking capabilities of its model (Methos), and refrained from launching it on a large scale.

However, a few days ago, Anthropic launched a version to the public called (Fable), which it said was equipped with electronic security protection measures.

Anthropic said in a blog post that the government informed it that it believes there is a way to bypass one of the protections that prevent the model from being used to find vulnerabilities that threaten cybersecurity.

The company added that the Trump administration had ordered it to prevent any foreign citizens, whether inside or outside the United States, from using its two latest models, which are Fable 5 and Mythos 5. In response, Anthropic said that it would disable access to the two models globally.

Amazon did not confirm whether it spoke to government officials about Anthropic’s models.