Amazon CEO Andy Jassy reportedly raised security concerns about Anthropic's AI models, leading to a U.S. government export ban on Friday. The Wall Street Journal reported Jassy informed Treasury Secretary Scott Bessent and other officials that Amazon researchers used Anthropic’s Claude Fable 5 to gather data potentially useful for cyberattacks, prompting a ban on the Fable 5 and Mythos 5 models.
Amazon's Role in the Decision
An Amazon spokesperson highlighted that while it’s common for governments to seek advice on security risks, the company does not disclose details of those discussions. The spokesperson also noted AWS's operations have been impacted by the model cutoff.
The Information and Reuters also reported on Amazon's security concerns related to Anthropic’s models. David Sacks, former AI czar under Trump and current co-chair of the President’s Council of Advisors on Science and Technology, indicated that a reliable source came forward with a jailbreak, leading to government action.
The Admin asked [Anthropic CEO Dario Amodei] to fix the jailbreak or de-deploy the model. Dario refused. — David Sacks
Government Response and Anthropic's Position
Sacks explained that the administration requested Anthropic CEO Dario Amodei to address the jailbreak issue or remove the model, but Amodei declined. Anthropic, in a blog post, argued that the capabilities causing government concern are available in other models that are publicly accessible.
Anthropic’s stance suggests that the potential threats perceived by the government might not be exclusive to their models. The company has not issued any further comments beyond the blog post.
Impact on Amazon and Anthropic
Amazon is a significant investor in Anthropic, and the ban affects its AWS operations. The extent of the impact on Amazon’s cloud services remains unclear. The developments raise questions about the future collaboration between Amazon and Anthropic in AI technology development.
This incident highlights the delicate balance between AI innovation and security. As AI models become more sophisticated, the potential for misuse grows, necessitating vigilance from both tech companies and government entities.
Source: https://techcrunch.com/2026/06/13/amazon-ceo-reportedly-raised-anthropic-model-concerns-before-government-crackdown/
