• Pentagon official Emil Michael warns that AI labs refusing to cooperate with government protocols could face export controls.
  • The White House's new AI task force has begun developing incident-reporting and cooperation requirements, with export restrictions as a potential enforcement tool.
  • The precise legal mechanism remains unclear, but officials cite temporary restrictions on Anthropic as precedent.

A Warning Shot for AI Labs

A senior Pentagon official is putting AI laboratories on notice: cooperation with the government’s new safety protocols is not optional, and refusal could trigger export controls. Emil Michael, the Pentagon’s undersecretary for research and engineering and vice chair of the White House’s Super Intelligence task force, made the warning in an interview with The Information, signaling that the administration is prepared to use trade restrictions as leverage over the frontier AI industry.

The task force, which adopted its charter on October 15, has launched a 120-day policy review covering cybersecurity, biosecurity, and model accountability. Michael will lead efforts to develop protocols for AI labs to report breaches, hacks, and other cyber incidents, and to coordinate engagement with frontier developers. The threat of export controls, according to a White House official who spoke to Semafor, is intended to ensure that labs comply with the voluntary accord that major developers—including OpenAI, Anthropic, Google (GOOGL), Meta (META), xAI, and Nvidia (NVDA)—have signed.

The precedent is clear: earlier this year, the Commerce Department imposed temporary restrictions on Anthropic after the company declined to provide the government with unfettered access to its models. Those restrictions were later lifted, but officials now point to that episode as proof that export controls can serve as a powerful enforcement tool. The exact scope and legal mechanism for such restrictions remain unresolved, and no new industry-wide ban has been enacted.

Industry Uncertainty

The warning adds a new layer of risk for an industry already navigating a complex regulatory landscape. AI companies, many of which derive significant revenue from international customers and partnerships, now face the possibility that their access to overseas markets could depend on their willingness to cooperate with Washington. For Anthropic, the stakes are particularly high: the company is preparing for a potential IPO, with investors reportedly valuing it near $1.5 trillion—down from earlier expectations of $2 trillion. The company’s third-quarter revenue reached $16.8 billion, according to an investor presentation described by The New York Times (NYT), and it remained profitable on an adjusted basis.

It is unclear whether the export-control threat has affected Anthropic’s IPO discussions. A spokesperson for the company declined to comment. The Pentagon did not respond to a request for comment by press time.

Meanwhile, the task force’s mandate extends beyond incident reporting. Vice chair Scott Kupor will work with the Labor Department to study AI’s impact on the workforce, while a Justice Department–FTC committee will examine consumer protection issues. The White House says it will consult civil-liberties advocates, consumer groups, and law enforcement as it develops its protocols.

The move toward coercive measures has drawn comparisons to China, which on October 16 proposed adding 40 technologies—including 23 AI-related ones—to its export control list. While the U.S. action focuses on lab cooperation rather than technology transfer, both signal that governments are increasingly treating AI as a strategic asset requiring state oversight.

What’s Next

The immediate question is whether the threat of export controls will be enough to bring reluctant labs to the table. The task force’s 120-day review will produce recommendations on cybersecurity, biosecurity, and model accountability, but it is not clear whether those recommendations will translate into binding rules. Officials have yet to publish a legal instrument or define specific reporting obligations. Until then, the warning is best understood as a credible escalation of pressure—not a new export regime.

For investors, the key takeaway is that government cooperation is becoming a commercial risk factor. The Anthropic precedent suggests that disputes with Washington can have consequences beyond federal procurement, potentially affecting overseas revenue and partnership opportunities. As the task force’s work progresses, the industry will be watching for the first enforcement action—or the first lab to test the government’s resolve.

Correction: An earlier version of this article misstated the timing of the task force’s charter adoption. It was adopted on October 15, not October 14.