President Donald Trump told the United Nations General Assembly on September 22 that the United States rejects efforts to build what he called a global system for controlling artificial intelligence. The remarks, reported independently by Scientific American and NBC 6 South Florida, put the U.S. president in open disagreement with U.N. Secretary-General António Guterres and other leaders calling for stronger safeguards and international cooperation.
The address is significant less as a new piece of policy than as a clear statement of diplomatic direction from a leading producer of AI systems. Trump did not announce a law, executive order, treaty withdrawal or binding U.S. regulatory action. Instead, he framed international control as a threat to technological growth, said he would not stifle AI, and argued that the field should be called “super intelligence.”
Safeguards divide the U.S. from U.N. leaders
Trump said the United States “totally rejects” attempts to create a global scheme to control AI, according to accounts of his speech. He also said the technology could become larger than the Industrial Revolution and perhaps the internet, claims presented as his assessment rather than independently established forecasts. A National News Desk report said Trump urged caution and said law enforcement could address problems if they arise. The speech did not define a legal mechanism, agency role or enforcement standard.
Guterres offered a contrasting view in remarks covered by NBC 6. He called for conditions that allow AI to develop responsibly and for added protections against its risks. He also said life-and-death decisions should not be surrendered to machines. Finnish President Alexander Stubb similarly called for controls that keep people in charge.
Trump did not identify a particular agreement he was rejecting. His remarks nevertheless opposed international construction of an AI-control framework, while Guterres and other leaders urged safeguards or cooperation.
Trump also asserted that the United States leads China and other countries in AI by a wide margin. The available reporting attributes that statement to the president; it does not provide evidence sufficient to measure a comparative lead across research, chips, model capability, investment, deployment or military use.
Existing White House order provides domestic context
The U.N. address does not establish that the administration has adopted no AI safeguards. The White House’s June 2 executive order, “Promoting Advanced Artificial Intelligence Innovation and Security”, directs work on AI-enabled cybersecurity and calls for capability benchmarking and a voluntary framework for certain advanced AI models.
The order also says that its relevant provisions do not authorize mandatory government licensing, preclearance or permitting for the development, publication, release or distribution of AI models. It therefore pairs security work and voluntary cooperation with developers with limits on federal gatekeeping for model releases.
The June order provides context for Trump’s comments, but he did not expressly invoke its domestic provisions in his U.N. speech. His remarks support the narrower conclusion that he opposes an international AI-control framework; they do not establish that the federal government has abandoned AI policy, cybersecurity measures or advanced-model evaluation.
“Super intelligence” is not a neutral replacement term
Trump’s proposed terminology imports a more specific concept into a broad public debate. “Artificial intelligence” covers a wide range of existing computational systems, from language models and image generators to tools used for classification, forecasting and automation. “Superintelligence,” by contrast, has an established technical and philosophical use for a hypothetical system whose cognitive capabilities exceed those of humans.
Calling all AI “super intelligence” blurs the distinction between current task-dependent systems and that speculative category. No documented directive has changed federal terminology, and other leaders at the U.N. continued to refer to artificial intelligence. For now, the proposed label is political rhetoric rather than an implemented policy standard.