This week the Nobel laureate, pioneering innovator, and chief executive officer of Google Deep Mind, Demis Hassabis, issued a rare public statement. He called for the creation of a U.S. led standards body to test “frontier AI models.” Clearly Hassabis is worried about the risks posed by rapidly advancing technologies. Just as clearly, given he is worried, so should we be.
Hassabis’s recommendation differed from other related ones. For example, he did not suggest an international regulatory body. Knowing the enormous practical and political difficulties involved in getting any sort of global cooperation, Hassabis proposed an initiative that was entirely American. Better to start such an effort in one country, he clearly concluded, than not to start one at all.
Along similar lines, knowing the politics, policies, and bureaucracies involved, Hassabis sought to avoid direct government (public sector) oversight. Instead, he proposed a self-regulatory body that would be funded by the tech industry (private sector), which would also oversee it. As one expert on the subject, Sebastian Mallaby, summarized, “Hassabis’s main point is that the self-regulatory route stands the best chance of generating a well-funded, technically proficient safety team without the world having to wait years for it.”
Implicit in Hassabis’s proposal were two overriding concerns: first, again, the risks posed by the new technologies. Second, the rank inability of government leaders – in the U.S. especially presidents and members of Congress – to tackle the problem on their own. Political leaders are incapable of addressing the issue both because of their generally profound differences and their generally profound ignorance of the technologies. Which is to say that if we leave it to our political leaders to lead us to a safe tech space it will not happen because it cannot happen.
The risks posed by the new technologies fall into two broad categories. One is from external sources, such as, to Americans, threats of competitiveness and control coming from China. And the other is from internal sources such threats to our social fabric posed by, for example, people replaced by robots leading to widespread unemployment.
To address the enormous problems posed by the advancing technologies there is no shortage of proposed solutions. A recent opinion piece in the New York Times suggested a “populist A.I. agenda that treats the technology as a public project.” It further suggested that we “treat some A.I. profits as a shared resource and distribute them directly to citizens.”*
Great. But… how do we get from here to there? The piece is full of good ideas, but there is nothing about how in the world they might be implemented. It’s one thing to have a good idea in theory; it’s something else altogether to put it into practice.
Which brings me back to where I started. To Hassabis’s suggestion that business leaders take the lead on this. That they do not wait for political leaders to guide us through a thicket about which most remain ignorant. That instead they, the experts, step to the front and point the way.
*https://www.nytimes.com/2026/07/11/opinion/ai-populism-china-open-source.html