A new report from The Wall Street Journal says the export control order that prompted Anthropic to cut off access to Fable 5 and Mythos 5 was driven in part by cybersecurity research from Amazon, as well as discussions between Amazon CEO Andy Jassy and the White House.
According to the Journal, Amazon produced a paper arguing that Fable 5 could be prompted into providing information that might be useful in carrying out cyberattacks. Amazon has not responded to a request for comment on the report.
Not long after Jassy shared those findings with federal officials, the government moved to restrict the models’ use by foreign nationals. That decision created immediate complications for Anthropic, where many researchers are foreign-born and were therefore blocked from using the company’s own systems.
Anthropic has pushed back on the government’s description of the issue, disputing the claim that the incident amounted to a “jailbreak.” The company said similar weaknesses could be identified in other publicly available AI models, including GPT 5.5.
Some security experts appear to agree with that view. Katie Moussouris, founder and CEO of LutaSecurity, wrote on BlueSky, “I’ve seen the paper. It’s not a jailbreak.” Former Commerce Department official Kate Koren also suggested to the Journal that White House hostility toward Anthropic may have played a role in the decision.
