Pentagon official floats export controls for AI labs that resist oversight, Semafor reports
Emil Michael, a Pentagon undersecretary who helps lead the White House AI task force, raised the idea in reporting published October 15
Published
Emil Michael, the Pentagon's Defense Undersecretary and a vice chair of the White House AI task force, has told colleagues the government could use export controls against artificial intelligence labs that will not cooperate with oversight, according to Semafor, which first reported the task force's work on October 15.
A White House official told Semafor that export controls were one option under discussion for labs found noncompliant. The same official pointed to a precedent: the Commerce Department had at one point temporarily imposed export controls on Anthropic after the company refused to give the government unfettered access to its models.
Michael sits on the task force known as the Super Intelligence Force, alongside chair Jay Clayton and members Andrew Ferguson and Scott Kupor. The group formally adopted its charter last Thursday, opening a 120-day window to work out how the government will handle oversight of artificial intelligence. At its first meeting, Semafor reported, how any noncompliance would actually be enforced, and how existing law might apply to companies that resist, was not yet settled.
The idea floated by Michael would be the first time export controls have been tied explicitly to a lab's cooperation with government oversight, rather than used as a one-off measure. It would also shift the task force from a body with undefined authority into one able to threaten enforcement, layering a new form of pressure on top of the earlier Anthropic case. Which labs would be judged uncooperative, which technologies would be restricted, and what would count as cooperation are all still undefined, leaving the real reach of the threat unclear.