A recent export control directive, which led to Anthropic restricting access to its Fable 5 and Mythos 5 models, was reportedly influenced by cybersecurity research from Amazon and direct discussions between Amazon CEO Andy Jassy and the White House. According to a report by The Wall Street Journal, Amazon's research paper claimed that specific prompting techniques could enable Fable 5 to generate information potentially useful for cyberattacks. Amazon has not yet provided a comment regarding these findings.
Following CEO Jassy's presentation of these findings to the government, a decision was made to prohibit the use of Anthropic's models by foreign nationals. This measure created a significant internal challenge for Anthropic, as many of its own researchers are foreign-born and consequently found themselves barred from accessing the very products they developed.
In response, Anthropic issued a statement challenging the government's characterization of the vulnerability as a "jailbreak." The company contended that similar security weaknesses could be exploited across various other publicly available models, including GPT 5.5. This perspective appears to be supported by some security experts. Katie Moussouris, founder and CEO of LutaSecurity, stated on BlueSky, "I’ve seen the paper. It’s not a jailbreak." Furthermore, former Commerce Department official Kate Koren suggested to The Wall Street Journal that the White House's perceived disapproval of Anthropic might have played a role in the directive.
This incident reignites a long-standing tension between Anthropic and the government. The company has previously been at odds with the Trump administration over its refusal to permit its AI technology for mass surveillance of American citizens or for powering lethal autonomous weapons. In February, then-President Trump instructed federal agencies to cease using Anthropic’s AI, and shortly thereafter, Secretary of Defense Pete Hegseth officially designated the company as a supply chain risk.
Despite these past conflicts, the government and Anthropic had recently appeared to mend their relationship, even collaborating to broaden access to Mythos. However, the latest directive suggests a renewed period of contention between the two parties.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.