Donald Trump has said AI needs only a “STRONG AND SMART (High IQ!) PRESIDENT” as its safeguard, while accusing Anthropic chief executive Dario Amodei of backing a “SICK conspiracy” against artificial intelligence and data centres.
In a post on Truth Social, Mr Trump argued that slowing the development of AI would benefit China. “WHOEVER WINS AI, WINS!” he wrote, as he urged Americans not to “kill the Golden Goose” – his description of the data centres being built to support the technology.
Mr Trump has warned that communities resisting data centres risk becoming “backwards and poor”, while the wider country benefits from jobs and investment. But opposition is widespread: 61% of Americans, including 54% of Republicans, said they would oppose a new data centre in their community in a poll by the Annenberg Public Policy Center at the University of Pennsylvania.
AI safety debate intensifies
The president’s intervention comes as leading technology executives call for greater caution over the development of increasingly powerful AI systems. Mr Amodei has urged the industry to “pace the frontier”, while OpenAI chief executive Sam Altman and SpaceX founder Elon Musk have also backed stronger safety measures.
OpenAI has called for mandatory national safety requirements based on the capabilities of AI systems. Mr Amodei has proposed independent evaluators, greater co-operation across the industry and deeper government oversight as the technology advances.
Mr Trump has rejected that direction. His vice-president, JD Vance, has described calls by AI laboratories to regulate themselves as a possible “Trojan horse” intended to secure regulatory capture.
The dispute has also unsettled investors. Nvidia and other chipmakers fell by more than 3% as markets considered whether the frontier laboratories might deliberately slow the development of new models, reducing demand for the vast infrastructure being built to support them.
Meta and Alphabet rose, while cybersecurity companies including CrowdStrike gained 15% or more. Gil Luria, head of technology research at D A Davidson, said regulation could favour the largest laboratories because they could afford extensive safety teams, lawyers and compliance staff.
He argued that rules developed with significant input from OpenAI and Anthropic could become a barrier protecting the companies already at the forefront of the industry.
Mr Trump said his administration already had “tremendous CRIMINAL and REGULATORY power over these companies”, while presenting restrictions on AI development as a possible advantage for Beijing.
The White House has become increasingly involved in the AI economy. A June executive order instructed national security agencies to develop benchmarks for the cyber capabilities of advanced models and establish when a system should be classed as a “covered frontier model”. It also proposed a voluntary framework allowing developers to provide the government with access to models before wider release.
The administration has separately clashed with Anthropic after the company restricted access to its advanced Fable and Mythos 5 models during a dispute over supplying the technology to the Department of War. Mr Trump has also floated taking stakes in AI companies, following similar intervention in industries including semiconductors and steel.
His position comes as Republicans face a narrow advantage in Congress ahead of November’s midterm elections. Democrats need three House seats to take control and four seats to change the balance in the Senate, while Mr Trump’s approval rating is described as being close to its lowest level of his presidency.
