King Charles held a private summit Thursday at Dumfries House, an 18th-century estate in Ayrshire owned by his charitable foundation, bringing together leaders from Nvidia, OpenAI, Anthropic, the U.K.’s new AI minister Kanishka Narayan, the head of the U.K.’s foreign intelligence service, and Paolo Benanti, who advises the pope on AI. The king made his position clear: he wants control over artificial intelligence before it’s too late.
The king urged attendees to find a way to control AI “before it’s too late,” calling the fast development of the technology and its “substance” both “intriguing and deeply concerning.” He cited medicine and science as examples of AI improving and saving life, but warned that creators are now increasingly warning AI risks developing darker capacities, perhaps even to take life.
An Anthropic researcher quit days earlier, writing online that AI companies were “racing to build machines that are much smarter than any human” and that humanity “may not survive this.” The king responded directly to that fear. He said it’s urgent to consider the existential dangers of AI falling into the wrong hands and being used in “potentially catastrophic ways.”
He asked, “Surely, we need sufficient means of control before it’s all too late?”
As he closed, he urged leaders to harness AI’s good and build international cooperation and consensus.
OpenAI sent CFO Sarah Friar and Anthropic sent Tino Cuéllar, who leads global affairs. Cuéllar said his company was honored to participate and grateful to the king for “convening industry leaders and voices from across society to discuss the role of AI in serving the public good.”
The summit follows Anthropic shutting off access to its Fable and Mythos models to only U.S. customers, which shocked the Euro tech scene. King Charles previously spoke about AI in 2023 and 2024.
This month, OpenAI’s Sam Altman, Elon Musk, and Anthropic’s Dario Amodei called for a deceleration in AI development. Nvidia’s Jensen Huang and President Donald Trump rejected those calls; the president said the U.S. is leading China in AI and wants to stay ahead.
OpenAI and Anthropic have teased blockbuster IPOs over the past year, though OpenAI recently said its IPO plans were stalled this year due to AI safety concerns. Outlets report the king stayed for around 20 minutes, then left guests to discuss among themselves.
The King’s Record on AI
Charles has spoken publicly about AI twice before. His language at Dumfries House was direct. He did not ban AI or shut down research. He asked for control, and he asked for it urgently.
The summit attendees included some of the most powerful figures in AI. The U.K.’s new AI minister was present, along with the head of the U.K.’s foreign intelligence service. Paolo Benanti’s presence adds weight to the argument that religion and science share responsibility for how AI develops.
What Comes Next
The summit was private and brief. The king stayed for around 20 minutes before leaving guests to talk among themselves. Whether that conversation produced any concrete steps is not yet known.
What is known is the direction of travel. The king has now spoken three times on AI, each time raising questions about control and risk. The summit format brought the people building the technology face to face with the person representing the state.
The question of whether the U.K. can hold its own in the AI race is a separate matter. The king’s call for control before it’s too late places him against the tech companies pushing faster development.
The king’s position is not anti-AI. He cited medicine and science as examples of AI improving and saving life. His concern is about control, not about the technology itself.
The Order of the Day
- The king’s warning: “Before it’s too late”
- The king’s assessment: AI development is “intriguing and deeply concerning”
- The king’s concern: AI could develop darker capacities, perhaps even to take life
- The king’s question: “Surely, we need sufficient means of control before it’s all too late?”
- The king’s closing call: Harness AI’s good and build international cooperation
The summit was a statement of intent. The king has spoken. Now the question is whether anyone listens.
Where the paper stands
The paper backs narrow rules against direct harm, such as forcing companies to disclose safety failures they hid, and is against federal licensing of AI that freezes today’s leaders in place and locks out whoever would have challenged them. The summit at Dumfries House was a statement of intent, not a blueprint for regulation. The king’s repeated warnings about control before it’s too late raise real questions, but they do not settle them.
The attendees — Nvidia, OpenAI, Anthropic, the U.K.’s new AI minister, the head of the U.K.’s foreign intelligence service, and Paolo Benanti — represent the people building the technology. Their companies are moving forward, with Anthropic pulling back access to its Fable and Mythos models to only U.S. customers. OpenAI’s CFO and Anthropic’s global-affairs lead were there, and the latter thanked the king for convening industry leaders and voices from across society.
The paper’s concern is the same as always: when the biggest firms ask to be regulated, the paper asks who those rules would lock out. Licensing regimes and compliance costs only giants can afford are a moat, not a safeguard. The summit did not propose a moat. It proposed urgency, control, and cooperation.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.