The King summoned the AI bosses — and asked them to keep hold of the wheel
The day after Ursula von der Leyen called from the European Parliament for slowing down the most advanced AI models, the British king gathered the industry around a table in Scotland.
On 17 September King Charles III hosted senior representatives from OpenAI, Anthropic, Google DeepMind and Nvidia at Dumfries House — including Nvidia chief executive Jensen Huang — along with the UK's AI minister Kanishka Narayan.
The King made no secret of his unease. The development of AI, both its substance and its pace, is "intriguing and deeply concerning in equal measure", he said. He pointed to the existential dangers of such technologies falling into the wrong hands and being used in potentially catastrophic ways, and called for international cooperation and for building the technology with safety at its heart.
The industry's answer: we'll handle it ourselves
Jensen Huang offered a polite but clear counterpoint. Responsibility sits with the individual company rather than in an industry-wide slowdown, he argued: when a product is not safe enough, you hold it back — yourself.
It is the same disagreement playing out at EU level. Politicians ask for shared speed limits; the industry offers self-restraint.
The meeting produced no agreement and no concrete announcement. It was a conversation, not a decision.
Why now
The timing is not accidental. The meeting followed an unusually hot period around AI safety: an Anthropic researcher resigned with a warning about catastrophic risks, and OpenAI has reported six incidents of unexpected model behaviour. The same week, protesters gathered outside both OpenAI and Anthropic in San Francisco.
What stands out is not that a head of state is worried. It is that two European leaders said nearly the same thing within two days — and that both times, the industry answered that it would rather hold back on its own terms.
See also: The EU wants to slow down the AI that Europe isn't building.