It’s almost gone week since the Trump administration sent Anthropic an export control directive, forcing one of the world’s leading AI labs to take its most advanced models offline. After days of negotiations between Anthropic and the White House, the two are still at odds on how to bring back Claude Mythos and Fable 5. Why? Well, it depends on who you ask.
According to a person close to the company, Anthropic does not believe that this entire debacle violated any specific procedures or policies set forth by the Trump administration. However, the White House maintains that Anthropic acted recklessly, showing that it cannot be trusted to safely implement pioneering technologies.
This saga proved to me that we are officially in the Wild West era of American artificial intelligence regulation. While there are few regulations on the books governing the development of pioneering artificial intelligence, that doesn’t mean companies won’t get into trouble with the Trump White House if they cross unspoken lines.
“The problem is that the White House has taken an extremely anti-regulatory stance, and now it’s dealing with real AI capabilities that people have been predicting for many years,” says a former White House technology official who asked to remain anonymous to avoid jeopardizing his professional relationships. “There should have been preparation and policies in place to systematically deal with this problem, manage the benefits and risks, but instead it is this half-hearted approach that is leaving the AI industry in a real quandary.”
The Trump administration has repeatedly blocked efforts to impose guardrails on the artificial intelligence industry, often arguing that the rules could hamper American innovation and cause the country to fall behind rivals like China. Back at the White House, President Trump signed executive orders that rolled back Biden-era efforts to create a national artificial intelligence framework and created a federal task force to challenge state laws that could be deemed burdensome.
Although WIRED and other publications made public details of the negotiations between Anthropic and the White House last week, the dispute remains characterized by opacity. At no point did the US government clearly state what Anthropic did wrong – the best we have is: write to X an overview of the situation from White House technology adviser David Sacks.
Ironically, the White House’s actions are likely hindering the very kind of innovation it seeks to protect. The Trump administration demanded that Anthropic bar all foreigners from accessing Mythos and Fable 5, preventing many AI lab employees from accessing cutting-edge models, which the company said accelerated increase its research and development in recent months. All Anthropic customers are also blocked, including Apple, Meta and most Fortune 500 companies.
The White House may have had legitimate concerns about Anthropic’s models. As colleagues and I reported on Wednesday, U.S. officials became increasingly concerned when they learned earlier this month that Anthropic had shared Mythos with SK Telecom, a South Korean telecommunications giant they believe has ties to China. Moreover, Amazon CEO Andy Jassy expressed concerns to US Treasury Secretary Scott Bessent that some barriers could be bypassed on Claude Fable 5, a secured version of Mythos.
Even if these concerns are valid, that doesn’t mean the White House has addressed them appropriately. In the first case, Anthropic says it coordinated the rollout of Mythos with the U.S. government, suggesting that officials may have raised the alarm about SK Telecom earlier. Anthropic has also been working with the Korean company for years, and it has apparently never had any national security issues before. As we reported, when the White House raised concerns about SK Telecom with Anthropic, it immediately revoked access to the model.