In a move that could reshape the global landscape of artificial intelligence governance, President Donald Trump is set to convene with the chief executives of America’s leading AI labs this week. The closed-door session, scheduled for Thursday in Washington, is expected to focus on unlocking private capital for domestic AI infrastructure while sidestepping the guardrails that European and British regulators have spent years constructing. For the United Kingdom, which has positioned itself as the world’s most credible AI safety steward, the development signals a profound geopolitical shift.
Sources familiar with the meeting indicate that the White House aims to fast-track data centre approvals, relax export controls on AI chips, and loosen federal oversight of large language model training. The initiative is being framed as an “AI Manhattan Project” to maintain US primacy over China. But critics warn this could trigger a regulatory race to the bottom, leaving Britain’s careful approach to AI risk management exposed.
The tension is most acute in the domain of frontier model safety. Britain’s AI Safety Institute, launched last November under Prime Minister Sunak, had hoped to set global standards for testing the most powerful systems. Yet with no mandatory compliance powers and a fraction of US private sector funding, its influence is waning. Meanwhile, Trump’s team has signalled it will favour voluntary commitments over legislation, echoing the industry-friendly posture that defined his first term.
“The UK bet on being the honest broker,” said Dr. Anya Kohli, a fellow at the Turing Institute. “But if the US decides to go it alone and carve out its own rules, the British approach becomes an irrelevant boutique standard. We risk a world where the only thing that matters is horsepower, not harm.”
For British tech leaders, the calculus is brutal. DeepMind, headquartered in London, now operates largely under US regulatory oversight. And without a domestic AI chip industry or a sovereign cloud provider, the UK is dependent on American infrastructure. A White House push to accelerate AI deployment could deepen that dependency, leaving Britain’s digital sovereignty in the rear-view mirror.
The European Union, for its part, has already passed the world’s first comprehensive AI Act, with mandatory compliance timelines starting in 2026. But Brussels’ approach is increasingly viewed as too slow and rigid for a sector moving at Moore’s Law pace. Trump’s deregulatory blitz could further isolate European and British firms from the global AI supply chain.
“This is a moment of choice for London,” said former Downing Street tech adviser Oliver Woods. “Either we double down on our safety-first model and accept that we might fall behind in AI capabilities, or we align with the US push for speed and risk becoming a junior partner in a less safe system.”
Thursday’s meeting is expected to include OpenAI’s Sam Altman, Google DeepMind’s Demis Hassabis, and Anthropic’s Dario Amodei. Also present will be venture capitalists from Silicon Valley’s most aggressive AI funds. The agenda is said to include recommendations for a national AI research cloud, streamlined visa processes for AI talent, and a joint industry-government task force on AI talent acquisition.
Critically, no British official has been invited. The UK’s AI Safety Summit last year drew global praise for its focus on catastrophic risk, but the conversation has now moved to Dubai, where the next global AI governance talks are set for February. Without a seat at the US investment table, Britain’s influence may be confined to the margins.
Proponents of the US approach argue that Europe and Britain are overreacting. “Regulation is a tax on innovation,” said venture capitalist Marc Andreessen in a recent podcast. “If we impose European-style rules on AI, we will cede the field to China. The only way to win is to build faster, better, and yes, with more compute.”
But the social costs of unbridled AI development are already apparent. Deepfakes are proliferating, algorithmic bias remains unchecked, and the environmental toll of massive data centres is mounting. A US push to accelerate AI could exacerbate these harms, especially if safety testing is relegated to a voluntary afterthought.
For the average Brit, the stakes may seem abstract. Yet the decisions made in Washington this week will shape the cost of healthcare, the reliability of autonomous vehicles, and the integrity of democratic discourse for years to come. The question is not whether AI will be governed, but by whom and for what purpose.
Britain still has a voice, but it must use it quickly. As one Downing Street insider put it: “We are no longer setting the pace. We are trying to keep up.”
