The Trump administration removed export restrictions on Anthropic’s Fable 5 and Mythos 5 artificial intelligence models Tuesday evening, ending a two-week ban and restoring customer access after weeks of uncertainty around the company’s most advanced products. The decision came from the top. Ordinary users, customers, and even allied governments had to wait while the state decided which machine-learning tools could move and which ones had to sit behind a gate.
Anthropic said access to Fable would return on Wednesday to users globally on several of its platforms, and it said it would reenable access on platforms such as Amazon’s AWS and Google Cloud as quickly as possible. The company also said it would continue to coordinate with the government to expand access to its more powerful Mythos model for select partners in the U.S. and elsewhere. That’s the arrangement now: access restored, but only after the company and the state finish their dance.
Who Has the Power
Commerce Secretary Howard Lutnick said in a letter to Anthropic that the company no longer required an export license for its products after agreeing to proactively detect and address security risks associated with the models, work with the government on protocols for future releases, and report any malicious activity it finds in its models. In a post on X, Lutnick said his office had worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the U.S. government and strengthen America’s leadership in AI. White House chief of staff Susie Wiles wrote on X that the government and private sector had worked together in a way we have never seen before and that this foundation of America First is unprecedented.
That’s the language of managed consent. The company gets access if it agrees to police itself, report on users, and build future releases around government protocols. The state gets a seat inside the product pipeline. Everyone else gets told this is progress.
Anthropic said it had imposed additional technical measures to block ill-intentioned users from bypassing some of Fable’s cyber guardrails. The company said Commerce’s Center for AI Standards and Innovation also tested the model’s previous and new safeguards. Anthropic said the new safeguard could cause Fable to flag more benign requests, but that in the other cases Fable’s output only included previously discovered or already patched security flaws. The apparatus calls that safety. The people using the tools get the friction.
Who Pays for the Rules
The administration issued export controls on June 12 that prohibited Anthropic from providing access to both Fable and Mythos models, then partially lifted the limits on Mythos late Friday before removing the restrictions on Fable Tuesday evening. Talks to lift the controls on Fable had been going on for over two weeks and continued over the weekend and into the week. The action came after months of friction between the administration and Anthropic, and amid a continued scramble within President Donald Trump’s inner circle to balance the cyber risks of advanced frontier AI models against the need to out-compete China.
That scramble landed on everyone below the decision-makers. The restrictions caused consternation throughout the AI industry and in Europe, where allied governments had been eager to use Anthropic’s products to find and plug vulnerabilities within their computer networks before adversaries do. The people doing the work had to live with the uncertainty while officials argued over leverage, risk, and national advantage.
The loosening of restrictions on Anthropic also came as OpenAI agreed to limit the availability of one of its most advanced models under pressure from the White House, adding to industry anxiety about the administration’s desire to regulate the technology. OpenAI’s ChatGPT-5.6 model was rolled out to a small group upon request from the White House, the Office of Science and Technology Policy and the Office of the National Cyber Director, despite complaints from CEO Sam Altman that this wasn’t how the company prefers to release its products. The release schedule, like the restrictions before it, ran through official channels first.
What They Call Coordination
Paul Lekas, head of global public policy and government affairs at the Software & Information Industry Association, said it was a positive development that Commerce had lifted the export controls on Mythos and Fable, but that there remained a real need for a consistent process and framework for frontier model assessment. Dean Ball, a former AI adviser in the Trump White House and the incoming head of strategic futures at OpenAI, said the White House’s reversal on the Anthropic controls only raised more questions. He said, “We have no idea what Anthropic did to make the models ‘safe,’ what commitments Anthropic has made going forward, and whether or how any of this applies to other frontier models in the government’s licensing queue.”
Joe Hoefer, chief AI officer of Monument Advocacy, said indefinite export controls on Anthropic’s products were never the most likely outcome and that the current outcome was more of a ceasefire than full resolution. Glenn Gerstell, the former general counsel of the National Security Agency from 2015 to 2020 and a senior adviser at the Center for Strategic & International Studies, said rescinding the export restrictions on Anthropic’s new models was the right step. He said, “We’re not going to maintain our lead over China in cutting-edge AI simply by slapping export controls on the latest innovation. The whole episode underscores how we have a lot more work to do to figure out the right way of balancing regulation for safety and national security reasons with promoting innovation in the AI sector.”
That’s the frame they keep returning to: safety, national security, innovation, leadership. The people at the bottom get the controls, the testing, the guardrails, and the uncertainty. The people at the top get to call it coordination and move on to the next licensing queue.