US President Donald Trump announces in a post on Truth Social from September 19, 2026, the establishment of a new federal agency called AI Force and intends to appoint a special AI envoy in the near future. He dismisses warnings from his own security research about out-of-control systems as exaggerated. He rejects binding regulations for the industry, stating that existing criminal and civil law should suffice.
Trump promises growth without new regulations
Trump compares the planned AI Force to the Space Force, which he established during his first term and which he believes has proven itself. He does not provide specific details about the agency’s tasks, budget, or placement within the government. According to Trump, only “high IQ individuals” should apply for the yet-to-be-named special AI envoy.
The government will neither hinder nor slow down the development of artificial intelligence but will promote and monitor it, Trump announces. Misconduct by companies can be punished under existing criminal and civil law, and he considers a separate AI regulation unnecessary. Just a few days earlier, he had referred to warnings about an AI takeover during a phone call with Nvidia CEO Jensen Huang as a hoax. Neither robots nor artificial intelligence would take control, he explained. He estimates the economic potential of the technology at up to 25 percent of U.S. economic output, along with the claim to maintain the lead over China. This forecast is the president’s own assessment and has not been independently verified.
Congressional representatives demand stricter rules
The announcement comes in a week where several congressional members are advocating for stricter regulations instead of a new agency without a clear mandate. Senator Bernie Sanders demands: “Stop the development of advanced AI. Ban artificial superintelligence. Immediately.” Republican representative Anna Paulina Luna calls for a special session of Congress on the topic – a bipartisan initiative that questions Trump’s wait-and-see approach within his own party.
Representatives Ted Lieu and Nathaniel Moran, a Democrat and a Republican, are also jointly introducing a bill. It would require developers to demonstrate that the system can be reliably shut down in an emergency before each model launch. Earlier that week, the House of Representatives had already passed a separate bill regarding the electricity costs of AI data centers with a vote of 417 to 3, albeit on a more narrowly defined topic. The clear outcome shows that Congress is significantly more capable of action on specific, narrowly defined AI issues than the previously content-free AI Force announcement from the White House would suggest.
The position of the special AI envoy has been vacant for months
Trump’s previous special envoy for AI and crypto, David Sacks, left the position in March 2026. He has since only been advising the government externally as the chairman of the President’s Council of Advisors on Science and Technology. This is also not the first attempt by the government to establish AI oversight. Already in July, the Treasury Department was working on its own AI oversight agency modeled after the financial regulator Finra. So far, no legal basis has emerged from this.
The current debate is also triggered by warnings from the industry itself. Former Anthropic researcher Jacob Coxon resigned in early September in protest against the race for superintelligence. Anthropic’s Alignment Chief Evan Hubinger subsequently estimated the risk of a fatal AI failure at over ten percent within ten years. Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have also recently publicly advocated for a slower pace of development, further highlighting the contrast to Trump’s course.
It will be crucial whether the announcement leads to an institution with actual powers or whether the AI Force, like previous oversight initiatives from the White House, remains stuck in implementation. A timeline, a budget, and a name for the special AI envoy are still missing – leaving the actual control question open.
