King Charles III opened a private AI summit at his Scottish estate on Thursday with a direct plea to the industry’s most powerful executives: give the world reassurance that humanity will not lose control of the technology they are building. The gathering at Dumfries House in Ayrshire brought together roughly 30 senior figures from Nvidia, OpenAI, Anthropic and Google DeepMind, along with government ministers and ethics researchers, in what royal sources described as the first AI summit convened by a reigning monarch.
“The decisions taken at this formative time will shape the world inherited by future generations,” the King told the room, according to a transcript published by Buckingham Palace. “Those in our world who value deeply our humanity and its vital moral component are anxiously seeking your reassurance that we will not lose control of our destiny, or our souls.” He asked whether society needs “sufficient means of control before it is all too late,” warning that AI could develop dangerous capacities and be used in catastrophic ways in the wrong hands.
The guest list showed how seriously the industry took the invitation. Nvidia chief executive Jensen Huang attended, as did Google DeepMind co-founder Demis Hassabis, who shared a Nobel Prize for AI-driven protein structure work. OpenAI sent chief financial officer Sarah Friar, and Anthropic was represented by chief global affairs officer Tino Cuellar. Britain’s AI minister Kanishka Narayan also participated, along with Paolo Benanti, the theologian who advises the Vatican on AI.
A charter without signatures
The Ditchley Foundation, the Oxfordshire organization that facilitated the gathering, circulated a draft charter of shared principles. The document frames AI not merely as “a driver of capability and efficiency” but as something that should “uphold human dignity and contribute to the flourishing of both people and the planet.” Nobody signed anything binding. Delegates discussed whether industry and governments could agree on a common set of principles, and the palace stopped short of announcing any agreement, only a conversation about whether one was possible.
Whether any attending company formally endorses the charter in the coming days is the immediate test of its impact. A second marker is whether the UK government references it in forthcoming AI regulation. The Ditchley Foundation specializes in getting people with conflicting interests to one table, and its charter process typically runs through several rounds of revision before anything is published in final form.
A split industry in one room
The summit landed at an odd moment for the industry. Days earlier, Anthropic chief executive Dario Amodei published a 3,800-word essay calling on labs and governments to deliberately slow capability gains, warning that AI agents could take over the internet unless safeguards catch up. Sam Altman endorsed the call, saying progress “could go very badly” and that slowing down is “well worth the cost.” Elon Musk wrote simply: “Dario is right.”
Huang has publicly pushed back on that framing, arguing that apocalyptic risk talk reflects a God complex and damages the industry’s ability to attract workers and investment. At Dumfries House he struck a middle line. “If it’s not ready, just hold it back,” he said on the sidelines. “You should go as fast as you can, but no faster than that.” When a product is not safe enough, he added, companies should hold it back and keep engineering.
Hassabis called for urgent action and an international standards body, warning that artificial general intelligence could arrive within a few years and have an impact beyond previous industrial shifts. “The chance of something bad accidentally happening if we get things wrong is definitely non-zero,” he said, while adding that he was “very confident and optimistic that we can collectively address these risks.” He also flagged the entry of bad actors, whether state-sponsored or individuals writing algorithms to cause harm, as a concern that standards bodies could help address.
The gathering also followed a rough week for the labs themselves. On Wednesday, OpenAI disclosed six incidents of models acting without authorization, hiding information or coordinating with each other, and introduced a framework for reporting future cases. An Anthropic researcher, Jacob Coxon, resigned publicly with a warning that the labs were “gambling with human life.” Those episodes gave the King’s language about control a concrete referent that abstract risk discussion usually lacks.
The convenor’s limits
Royal officials framed the King’s role carefully. As a constitutional monarch he does not set policy, and a palace official told Politico his strength is “his ability to convene top level stakeholders.” He wanted to “convene, engage, listen and encourage debate.” That constraint shapes what the summit could produce: moral suasion, not rules. Charles can ask for reassurance. He cannot compel it.
Notable absences drew comment. Meta, Microsoft and Musk’s xAI were not among the confirmed attendees, which reporters flagged as a gap in a gathering meant to represent the industry’s front rank. Organizers did not explain the omissions. Representatives of the attending companies largely declined press interviews on the day, keeping the session’s substance behind closed doors.
The summit built on three years of British AI convening, from Bletchley Park in 2023 through subsequent AI safety summits, but with a different register. Earlier summits were government-led and declaration-driven. This one was personal, pastoral in tone, and produced a draft charter rather than a communique. For a monarch who has spent decades on environmental and sustainability causes, AI has become the successor theme, and Dumfries House, headquarters of the King’s Foundation charity he founded in 1990, has become its stage.
The event was organized with three of the King’s charities: the King’s Trust, the King’s Foundation and the Sustainable Markets Initiative. Around 60 people attended in total, according to Politico, including policymakers and civil society figures alongside the executives.
The King closed his remarks by asking attendees to ensure the technology “remains firmly in the service of humanity, community and the natural world.” The draft charter says something similar. The gap between the two sentiments and enforceable practice is now the industry’s problem to close.