Multiple reports say Amazon CEO Andy Jassy raised security concerns with US government officials that contributed to the government blocking customer access to Anthropic’s advanced AI models. The models identified are Anthropic’s “Fable 5” and “Mythos 5.” According to the accounts, Jassy told officials, including US Treasury Secretary Scott Bessent, that Amazon research indicated the models could be used to extract sensitive information in ways that might support cyberattacks. The reporting also describes the government’s action as a ban on foreign access or customer availability of these models. Anthropic disputes the characterization of the risk, stating that the vulnerabilities Amazon allegedly highlighted are minor and already publicly known, according to the summary provided by one outlet. The combined coverage frames the episode as stemming from government engagement with major AI companies over security and access restrictions, with the central point of disagreement focused on the severity and novelty of the alleged weaknesses.