The White House Is Making Its AI Rules In Real Time

It has been almost a week after the Trump administration sent it export control instructions to Anthropic, forcing one of the world’s leading AI labs to pull its most advanced models online. After days of discussion between Anthropic and White House, the two would still be at odds over how to bring the Claude Mythos to Fable 5. Why? Well, it depends on who you ask.

Throughout the process, Anthropic is not believed to have violated policies or regulations set forth by Trump, according to a person close to the company. But the White House argues that Anthropic acted recklessly, suggesting that it cannot be trusted to produce borderline technology.

This saga convinced me that we are now in the Wild West era of American AI law. While there are few laws on the books governing AI development, that doesn’t mean companies won’t be in trouble with Trump’s White House when they cross unspoken lines.

“The problem here is that the White House has been very vocal about these rules, and now they’re facing the real power of AI that people have been predicting for years,” said a former White House technology official, who asked not to be identified so as not to jeopardize their working relationship. “There should have been planning and policies to deal with this, to manage the pros and cons, but instead it is this approach that puts the AI ​​industry at real risk.”

The Trump administration has repeatedly blocked efforts to compel the AI ​​industry, often arguing that the rules could undermine US innovation and cause the country to fall behind rivals like China. Since returning to the White House, President Trump has signed major executive orders change it the Biden-era effort to create a national AI system and create an army to challenge the government’s own laws can be considered difficult.

Although WIRED and other publications have detailed the discussions between Anthropic and the White House publicly over the past week, the dispute is still defined by its ambiguity. Nowhere has the US government clearly stated what Anthropic did wrong – the best we have is post on X explaining what is going on at the White House adviser David Sacks.

Ironically, what the White House is doing has thwarted the very nation it seeks to protect. The Trump administration wanted Anthropic to ban all foreign nationals from entering Mythos and Fable 5, preventing many of the AI ​​lab’s employees from accessing its most advanced models, which the company says has advanced. expanding its research and development in recent months. All of Anthropic’s customers are closed, too, including Apple, Meta, and many Fortune 500 companies.

The White House may have had good reason to be concerned about Anthropic models. As my friends and I reported on Wednesday, US officials were worried after hearing about it earlier this month Anthropic shared Mythos with SK Telecoma South Korean telecom giant that is said to have ties to China. Personally, Amazon CEO Andy Jassy he raised concerns to US Treasury Secretary Scott Bessent that some of the guards of Claude Fable 5, the protected version of Mythos, can be avoided.

While these concerns are legitimate, it does not mean that the White House has adequately addressed them. For starters, Anthropic is said to have cooperated with the US government in the release of Mythos, meaning that there would be a chance that authorities would be informed about SK Telecom ahead of time. Anthropic has also been working with the Korean company for years, and the system appears to have never run into any national security issues. And when the White House raised concerns about SK Telecom’s move to Anthropic, it terminated access to the brand immediately, we reported.