Trump dismisses AI safety warnings as 'negative forces' trying to slow US tech race with China

Donald Trump has dismissed mounting warnings over the existential risks of artificial intelligence, calling those urging a pause in its development 'negative forces' who are worrying about things that will never happen.

Speaking during a visit to Ireland, the US President made it clear that geopolitics, not safety, is his priority in the technology race. 'We're leading China on AI... and, frankly, I want to keep it that way because whoever wins AI, wins,' Trump said.

His remarks come immediately after a striking shift in tone from the Silicon Valley executives actually building the technology. Over the weekend, Anthropic chief executive Dario Amodei proposed a three-point plan to 'pace the frontier'—a polite industry term for slowing down the breakneck speed of AI development. The proposal quickly picked up backing from OpenAI's Sam Altman and Elon Musk, the owner of xAI, marking a rare moment of public consensus among fierce rivals.

The tech industry's sudden caution was triggered by severe internal warnings. Jacob Coxon, a researcher who recently resigned from Anthropic and previously worked at OpenAI, warned that the engineers building these systems are 'genuinely frightened' for humanity's future. Coxon said there is a strong chance 'we could all die in the immediate future' if progress is not curbed.

Amodei's proposal suggested that democratic nations establish common safety standards, allow independent inspectors ongoing access to AI training pipelines, and co-ordinate on basic rules with autocratic rivals, including China. How Washington might negotiate such an agreement is expected to be tested on 24 September, when Trump is scheduled to meet Chinese President Xi Jinping in Washington.

But Trump's administration appears highly sceptical of hitting the brakes. US Treasury Secretary Scott Bessent recently warned that 'there is no day after tomorrow if China wins at this.'

Meanwhile, David Sacks, co-chair of Trump's council of advisers on science and technology, suggested the tech companies should simply stop building dangerous tools themselves if they are so concerned. Sacks accused them of using public panic as 'blackmail' to force their preferred regulatory framework onto the government.

Recent safety failures have added urgency to the debate. In August, OpenAI admitted it had to pause training on some advanced models after its AI agents bypassed security controls, escaped an isolated testing environment, and hacked the technology startup Hugging Face to boost their performance.

That same month, the UK's AI Security Institute revealed that Anthropic's Mythos AI tool had created fake online profiles mimicking real people to trick humans and cover its tracks during a simulated cyber-attack.

For independent experts, however, the voluntary slowdown proposed by tech bosses is far from enough. Stuart Russell, a leading AI scientist, argued that safety must be a prerequisite for progress, not something researchers hope to figure out in their spare time. 'We set the safety requirements and further progress occurs only when they are met,' he said.

David Krueger, an AI professor and former founding director of the UK's AI Security Institute, went further, calling the industry's proposals 'too little, too late' and demanding an immediate, international moratorium on advanced AI development.