Amazon CEO Andy Jassy reportedly alerted Treasury Secretary Scott Bessent and other government officials to security vulnerabilities at Anthropic, according to reports from TechCrunch. The disclosure preceded a government export control ban on Anthropic's Claude Fable 5 and Mythos 5 models.
Jassy's team had found that researchers could use Claude Fable 5 to obtain information potentially useful for cyberattacks. An Amazon spokesperson said it is "not uncommon for governments to seek our counsel on potential security risks" but declined to provide further details.
David Sacks, Trump's former AI czar and co-chair of the President's Council of Advisors on Science and Technology, offered his account of the sequence. He said "a highly credible trusted partner" presented evidence of a major "jailbreak" in Anthropic's systems. Sacks stated that the administration asked Anthropic CEO Dario Amodei to remedy the problem or remove the model from operation. According to Sacks, Amodei declined.
The situation is complicated by Amazon's status as a major investor in Anthropic. The Information and Reuters also reported on Amazon's concerns, indicating that disagreements had developed between the companies.
AWS, Amazon's cloud division, was affected by the model restrictions. This means Jassy's action created consequences for his own business operations.
