The U.S. Defense Department's chief technology officer opposed the government acquiring stakes in AI companies and nationalisation. He said corporate responsibility for safety and stronger enforcement of existing laws should be expanded rather than widening new regulation.

On Sept. 16, CNBC reported that Emil Michael (에밀 마이클), the Pentagon CTO, answered "I hope not" when asked about the possibility of the government holding a partial stake in AI companies. Michael said he did not want direct government intervention in the AI industry and also expressed a negative view of nationalisation.

The comments differ from the Trump administration's past moves to expand equity investments in private companies. The U.S. government has secured a 10 percent stake in Intel and holds a so-called golden share in U.S. Steel, a unit of Japan's Nippon Steel, giving it a say in certain management decisions.

Michael also took a cautious stance on expanding AI regulation. He argued for an approach that holds companies responsible for developing products that are safe and aligned with human intent, and for releasing them only after meeting safety requirements, rather than European-style pre-emptive regulation.

On the recent Hugging Face hacking incident, he acknowledged concerns but said it was unclear what regulation could have prevented it. He also said it was risky to leave all judgments to executives at a small number of AI companies, and stressed that he would enforce existing laws, including those related to the Federal Trade Commission (FTC).

Michael also said there was an organised campaign behind the push to highlight AI risks, aimed at inducing decisions favourable to some established companies. He added that companies could voluntarily pause development of some technologies until they address safety problems.

Michael's remarks align with the Trump administration's policy stance of supporting industry growth due to AI competition with China while opposing limits on the pace of development. It remains debated whether corporate voluntary safety measures and existing laws alone can sufficiently respond to AI risks.