The U.S. government lifted export controls on Anthropic's Fable 5 and Mythos 5 models on June 30, reversing a freeze that had blocked access for weeks. The order had followed a cybersecurity finding that triggered the initial export restriction.
The move has drawn sharp criticism from security experts, including Alex Stamos, who called the now-reversed controls an "own goal" that creates uncertainty. Stamos argued the restrictions could push service providers toward Chinese AI models instead of American ones.
Anthropic confirmed that none of the jailbreaks cited in the cybersecurity finding offered capabilities beyond what other models—including Chinese ones—could already achieve. The models were restored for all users after the clearance.
The Trump administration's approach has been described as erratic, leaving companies across the industry with little clarity on future governance. The brief freeze highlighted how export policy can shift rapidly based on security assessments.
Critics warn that such reversals undermine U.S. credibility and incentivize reliance on non-American AI providers. The episode underscores tension between security concerns and maintaining competitive advantage in the global AI market.