Amazon security research reportedly led to the White House’s Anthropic Fable ban

Amazon’s cybersecurity team uncovered vulnerabilities in Anthropic’s Fable 5 and Mythos models that could enable cyberattacks, according to a *Wall Street Journal* report. The findings were shared with U.S. officials, including White House discussions with Amazon CEO Andy Jassy, prompting the government to impose export controls. These restrictions now block foreign nationals—including many of Anthropic’s own researchers—from accessing the models. Anthropic rejected the government’s framing of the issue as a 'jailbreak,' stating that similar vulnerabilities exist in other publicly available AI systems, such as GPT 5.5. Security experts like Katie Moussouris, CEO of LutaSecurity, agreed, calling the paper’s claims overstated. Meanwhile, former Commerce Department official Kate Koren suggested the White House’s preexisting hostility toward Anthropic may have influenced the decision. The latest clash follows a long-standing dispute between Anthropic and the Trump administration. The company has refused to allow its AI to be used for mass surveillance of Americans or autonomous weapons, leading Trump to order federal agencies to stop using Anthropic’s technology in February. Hours later, Secretary of Defense Pete Hegseth designated Anthropic a supply chain risk. Earlier this year, relations appeared to improve as the White House and Anthropic collaborated to expand access to Mythos through Project Glasswing. However, the new export controls signal a return to conflict, raising concerns about the future of U.S.-AI company cooperation. Amazon has not yet responded to requests for comment on the research or its role in the decision. The export restrictions now limit global access to Anthropic’s advanced models, complicating research and development for the company’s international team.