Trump calls AI guardrail talk a conspiracy and says a president is enough

President Donald Trump wrote on Truth Social on 14 September that calls for new artificial-intelligence limits are a "SICK conspiracy" against AI and data centres, and that "the only one that is happy about it is China." The only control the industry needed, he said, was "a STRONG AND SMART (High IQ!) PRESIDENT," which the United States already had. He added that his administration had already stopped AI people from doing bad or potentially bad things, named Anthropic chief executive Dario Amodei as someone now "pretending to be a 'perfect little angel'," and warned "Conspiracy Theorists, Treasonists, Traitors, and Leakers."

The post was the second day in a row in which Trump dismissed a slowdown. Amodei published a long essay on Saturday arguing that the industry should ease the pace of capability gains so that limits can be built. He wrote that in six to 12 months, AI agents could be capable of taking over the entire internet and causing hundreds of billions of dollars in damage. Other industry figures, including Sam Altman, Elon Musk and Demis Hassabis, have in recent days also spoken in public about slower deployment or tighter oversight. UN High Commissioner for Human Rights Volker Turk sent a letter dated 14 September to ten labs, including Anthropic, OpenAI, Google, Meta and Microsoft, saying industry self-regulation is not enough.

What Trump says already exists

Trump wrote that Washington already holds "tremendous CRIMINAL and REGULATORY power over these companies" and will keep using it. He did not list new statutes. He did not answer Amodei's six-to-twelve-month timeline. He framed the argument as a race: "WHOEVER WINS AI, WINS!" He said the United States is leading China and will continue to do so. Data-centre siting, which has become a midterm issue in several states because of land, water and power use, was folded into the same conspiracy sentence rather than treated as a local planning dispute.

That packing is the useful part of the post. One message tries to cover three separate fights: lab safety, Chinese competition, and county-level resistance to server farms. Lab safety is what Amodei and Turk are writing about. Chinese competition is a trade and export-control problem. Data-centre opposition is zoning, electricity prices and groundwater. Treating all three as a single plot against the industry lets the White House reject new federal rules without taking each fight on its own terms.

The week around the post

The UN Security Council is preparing a session on AI. Senate Republican leader John Thune has said he sees a role for Congress and has been working with Commerce Chair Ted Cruz on a bill. Trump's line is that no such bill is required. China's own security chief, Chen Yixin, wrote in China Cyberspace on 13 September that foreign models could be used for espionage and strikes on infrastructure and that the Party needed tighter control. Beijing's public diplomacy still tells Washington not to hype AI risk. Its internal security writing does the opposite.

Trump's text is not a regulatory filing. It is an instruction to his own side and a warning to anyone inside the administration or the labs who wants slower release. The practical test comes if Congress does bring a bill, or if a model incident lands before November. Until then the official US position, as stated by the president on 14 September, is that existing criminal and regulatory tools plus the occupant of the office are the guardrail, and that asking for more helps China.