US President Donald Trump said on September 19 that his administration would create an “AI corps” and appoint a director for artificial intelligence, signaling support for faster industry expansion even as concerns grow over safety risks and the rapid build-out of data centers.
In a post on Truth Social, Trump said the United States would not “in any way impede or slow down” the AI industry. Instead, he said the government would “cherish, help and supervise” the sector. He dismissed calls to slow AI research as a politically motivated “hoax.”
Trump outlines plan for an “AI corps”
Trump said the proposed AI corps would follow a model similar to the US Space Force, whose creation he supported during his first term. He described the Space Force as a “tremendous success,” but did not provide details on the new body’s staffing, budget, leadership structure or legal authority.
Trump also indicated that the administration would address misconduct by AI companies through the existing criminal and civil justice systems. That suggests the AI corps is currently more of a policy and administrative proposal than a defined enforcement agency. It remains unclear whether its role would involve regulatory coordination, technical safety, law-enforcement support or another function.
Trump said AI could eventually account for 25% of US gross domestic product and have a greater impact than the internet. He said the United States currently leads China and other countries in the field and wants to preserve that advantage.
Data-center construction has become a significant issue during the US midterm election cycle. Trump said opposition to new data centers had largely failed, but provided no supporting figures. The facilities’ power demand, pressure on local infrastructure and environmental effects continue to draw opposition in several areas.
White House AI policy points in different directions
The Trump administration has previously taken steps that placed limits on AI companies. The Commerce Department temporarily imposed export restrictions on Anthropic’s most advanced AI models. The Defense Department also sought to designate Anthropic as a supply-chain risk, but a judge later ruled that move unlawful.
David Sacks initially served as the administration’s AI and cryptocurrency czar early in Trump’s second term. He later left that position to become co-chair of the President’s Council of Advisors on Science and Technology. Trump’s proposal to appoint a new AI director suggests the White House may be revisiting how it coordinates AI policy, although no candidate has been named.
The administration’s support for industry expansion contrasts with recent discussions among company executives about slowing research. As AI systems become more capable, concerns within the sector have intensified over loss of control, misuse and inadequate safety testing.
Anthropic and OpenAI executives debate safeguards
Anthropic Chief Executive Dario Amodei previously said the company would give independent evaluators long-term internal access similar to employee privileges, allowing them to assess the safety performance of AI systems on an ongoing basis. Amodei described the arrangement as part of a broader safety program and said the pace of AI research should be reduced.
OpenAI Chief Executive Sam Altman, SpaceXAI Chief Executive Elon Musk and Google DeepMind co-founder Demis Hassabis also agreed to take part in related discussions. Altman said OpenAI and other leading AI laboratories could be close to an agreement to slow development and jointly address growing safety risks.
Anthropic researcher Jacob Coxon previously announced his resignation, accusing Anthropic and his former employer, OpenAI, of taking an insufficiently cautious approach to developing more powerful AI systems. In a post on X, Coxon wrote that the companies were moving directly toward self-improving superintelligence and were betting human lives on the outcome.
Evan Hubinger, Anthropic’s head of alignment science, responded that many researchers at Anthropic and OpenAI genuinely believe increasingly capable AI could lead to human extinction. The comments show that views on long-term risks remain divided within the industry. Those risks also involve unresolved assumptions about technology, governance structures and evaluation methods.
Antitrust lawsuit challenges proposed slowdown agreement
The debate over slowing AI development has also prompted a new antitrust lawsuit. Filed on September 18 in the US District Court for the Northern District of California, the case accuses Anthropic, OpenAI, SpaceXAI and Google of coordinating to slow AI development in violation of antitrust law. The plaintiffs argue that such an arrangement could reduce the value of AI subscription services they have already paid for.
The plaintiffs said they do not oppose AI companies asking the federal government to establish rules or seeking antitrust exemptions. Their lead attorney, Nick Rowley, said that if AI safety standards and related agreements are controlled through private arrangements by the world’s most influential for-profit technology companies, AI could quickly move beyond human control and produce severe consequences.
The lawsuit contains allegations by the plaintiffs, and the court has not made a final determination on whether the companies violated antitrust law. For AI companies, the challenge of building verifiable safety mechanisms while competing for computing capacity, talent and market share will affect research timelines, regulatory relationships, the value of subscription products and potential legal exposure.