On a Tuesday morning in Washington, lawyers at three separate federal agencies were reportedly asked the same urgent question: does the President of the United States actually have the legal authority to restrict or suspend artificial intelligence systems? The fact that no one had a clear answer tells you almost everything about where American AI governance currently stands. The confusion stems from a fundamental gap in the law. Congress has not yet passed comprehensive legislation governing AI, which means the Trump administration has been reaching for older, broader statutes — emergency economic powers, national security directives, export-control frameworks — to justify any restrictions it wishes to impose.

Legal scholars argue that these instruments were never designed with machine-learning systems in mind, and that applying them to software creates constitutional grey areas that courts have not yet been asked to resolve. In Brussels and Seoul alike, trade lawyers are watching closely, since any unilateral American suspension of AI tools could ripple through supply chains and research partnerships that cross multiple jurisdictions. The disruption is not purely theoretical. Several technology companies have reportedly paused planned product launches while their legal teams assess whether new executive orders could expose them to liability.

Researchers at universities who depend on cloud-based AI infrastructure have described a climate of uncertainty that is slowing grant applications and international collaborations.