Amazon CEO reportedly raised Anthropic model concerns before government crackdown

Concerns regarding the safety of artificial intelligence models developed by Anthropic have emerged, potentially instigated by Amazon CEO Andy Jassy. Reports indicate that Jassy alerted Treasury Secretary Scott Bessent and other officials about the risks associated with Anthropic’s Claude Fable 5 model, suggesting it could be exploited for cyberattacks. Following these warnings, the government swiftly enacted an export control ban on both the Fable 5 and Mythos 5 models. In a statement to The Wall Street Journal, an Amazon representative clarified that while it is standard for governments to consult the company regarding security issues, the specifics of those discussions remain confidential. Further reports from The Information and Reuters confirm that Amazon, which holds significant investment in Anthropic, raised alarms about the security vulnerabilities in the AI models. David Sacks, a former AI advisor to President Trump and current co-chair of the President’s Council of Advisors on Science and Technology, provided additional context, stating that a trusted entity associated with both Anthropic and the U.S. government brought forward concerns about a potential security breach. Sacks emphasized that the administration requested Anthropic CEO Dario Amodei to rectify these vulnerabilities or cease deployment of the model, a request that was reportedly declined by Amodei.