Amazon voiced concerns about Anthropic AI models before US crackdown, source says​

• The US (Trump administration) ordered Anthropic to block foreign nationals from using its advanced Fable 5 and Mythos 5 AI models due to security risks, including a "jailbreaking" method (Anthropic blog post).
• In response to these US export controls, Anthropic globally disabled access to its Fable 5 and Mythos 5 AI models for all foreign nationals (Reuters).
• Experts criticised the broad US ban on Anthropic's AI, calling it "not well thought-out" by the Trump administration as it affects allied nations' researchers.

WASHINGTON – Amazon chief executive officer Andy Jassy was among tech leaders who raised concerns to senior US administration officials this week about security risks in Anthropic’s most advanced AI models, a person familiar with the matter told Reuters.

Jassy’s involvement sheds light on the extraordinary move by Anthropic on June 12 to shut down its latest models globally in response to national security orders from the administration.

The San Francisco-based AI startup, which has confidentially filed for a US initial public offering, had previously warned about the hacking capabilities of its Mythos model and held it back from wide release. But earlier this week, Anthropic rolled out a public version, called Fable, with what it described as cybersecurity safeguards.

That brief release ended on June 12. In a blog post, Anthropic said the US government told the company it believes there is a method of bypassing, or “jailbreaking”, a safeguard against using the model to find cybersecurity holes.

The bypass found only “minor” security flaws that other publicly available models can also find, Anthropic said in its blog post.

The government ordered Anthropic to block any foreign nationals, whether inside or outside the US, from using both its latest models, Fable 5 and Mythos 5, the company said. In response, Anthropic said it would disable access to the models globally.

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

“As a leading cloud provider that serves a large number of private and public sector customers, it’s not uncommon for governments to seek our counsel on potential security risks,” an Amazon spokesperson said. “When they occur, we don’t share details of these discussions.”

Export control

The Information, a technology news outlet, reported Jassy’s concerns on June 13. Citing a US official, it later reported that the administration was unlikely to force other AI firms to abide by restrictions similar to those placed on Anthropic.

Government restrictions came in the form of an export control, Anthropic said in its blog post.

The US Commerce Department’s Bureau of Industry and Security, which oversees export controls, did not immediately respond to a request for comment.

Officials issued the export control “reluctantly” after Anthropic CEO Dario Amodei “refused” to “fix the jail break or de-deploy the model”, White House adviser David Sacks wrote in a social media post on June 13.

“The Admin’s hope is that Anthropic remediates the safety issue, the export control is lifted, and Fable goes back into general release,” wrote Sacks, co-chair of the Council of Advisors on ​Science and Technology and formerly the White House’s AI czar.

Some experts who favour export controls on advanced AI models found the government’s action puzzling because it affects allied nations as well as adversaries.

“This was not well thought-out,” said Jimmy Goodrich, a senior fellow at the University of California’s Institute for Global Conflict and Cooperation. “It even bans Canadians and Brits employed at Anthropic from doing research and development.”

The order came just as a previous dispute between government officials and Anthropic, which has confidentially filed for a US initial public offering, showed signs of easing. REUTERS