It is a sentence that would have read as satire five years ago. King Charles III waded into the artificial-intelligence debate on Thursday, gathering top technology executives in Scotland to urge them to hash out guidelines to keep the technology from spinning out of control. In the room: Nvidia chief executive Jensen Huang, Alphabet chief scientist Demis Hassabis, executives from Anthropic and OpenAI, an AI adviser to the Vatican, and the chair of China’s expert committee for AI governance.
“Those in our world who value our humanity and its vital moral component are anxiously seeking your reassurance that we will not lose control of our destiny,” the king told them, according to remarks released by the palace.
Why a monarch
Max Colchester and Jenny Strasburg explain the strategy in Friday’s Journal, and it is more calculated than it looks. Britain is not a superpower in this technology and has no realistic prospect of becoming one. What it has is convening power, and a genuine cluster — Google DeepMind’s headquarters, Oxford and Cambridge nearby, an OpenAI research hub committed to London. It holds the G-20 presidency next year. Deploying the king is part of an attempt to be the honest broker between Washington and Beijing.
“Charles is bringing the voice of reason into what has become a very inflammatory, polarized debate,” said Prof. Alan Woodward of the Center for Cyber Security at the University of Surrey, who adds that the goal is less about debating whether AI will wipe out humans than about making sure other countries don’t cede governance entirely to the U.S.
Britain is not alone in trying. European Commission President Ursula von der Leyen said this week she will invite AI companies to consider a slowdown. Pope Leo published an encyclical earlier this year warning that AI threatened to normalize an “anti-human vision.”
The uncomfortable footnote
The U.K. created what is now the AI Security Institute in 2023 to evaluate advanced models independently before public release. Other countries, including the U.S., Canada and Singapore, built their own.
And then the institute told U.K. lawmakers this week that Anthropic’s Mythos 5.1 model was not made available for pretesting to it, or to any AI safety agency outside the United States. The institute did not say whether it had been shown the results of testing by its American counterpart.
That single fact is the whole argument about sovereignty in one line. An institution built specifically to evaluate frontier models was not given the frontier model.
Nick Bostrom — the philosopher whose book shaped these concerns, and which a young Jacob Coxon read at school, as the Journal reports elsewhere in the same edition — puts the risk for everyone else bluntly: “A big risk if you’re not China or the U.S. is that you will be cut off from access to the most advanced AI at some point,” which could leave a country “instantaneously economically uncompetitive and geopolitically completely vulnerable.”
Our read
There is a portfolio observation here and it is not about which model wins.
If access to frontier capability becomes a geopolitical instrument — something granted, withheld or licensed rather than simply bought — then the economic value of AI concentrates in fewer jurisdictions and fewer companies than the current investment narrative assumes. That is a reason to hold this theme through the largest, most diversified platforms rather than through any single national champion or any pure-play supplier, which is how it is held here: Microsoft (MSFT), Alphabet (GOOGL) and Meta (META) alongside the grid and industrial builders that get paid regardless of whose model wins.
It is also a reason to treat every confident timeline about regulation as entertainment. A king, a pope, a European Commission president and a U.S. Congress that just went home without voting are all currently attempting to shape the same technology. The only safe prediction is that the outcome will not resemble anybody’s slide deck.