The most recent G7 summit included the leaders of seven of the most powerful countries in the world — and executives from OpenAI, Anthropic and DeepMind, who were treated like peers. AI companies might seem like they’re fighting over only product and market share, but they’re playing a much bigger game: This is a struggle for power at the highest levels.
AI companies aren’t shy about the scale of their ambitions. The leading labs all say they are building artificial general intelligence. Even if you don’t think they’ll succeed, their willingness to spend hundreds of billions of dollars in the attempt is telling. The industry’s leaders issue warnings about the dangers of a power-hungry runaway AI that tries to take over the world. Their actions suggest that they understand that motivation extraordinarily well.
No company is monopolizing AI yet, but that’s not necessarily reassuring. The valuations on these companies only make sense if one of them ends up a monopoly, and a company that monopolized AI might be the most powerful one in history. If none does, that means they will be competing for power with each other as well as the state. That could make the dynamics more intense, not less.
Microsoft Chief Executive Officer Satya Nadella warned recently that “there is no societal permission” for a future in which the profits from AI are concentrated in a handful of companies. Nadella was describing the economic danger. The deeper one is not who captures the profits, but who captures the power.
We’ve seen companies with the power of nations before. The British East India Company was chartered in 1600 and became a private corporation that governed an empire. It minted coin, levied taxes, ran courts and fielded an army of some 260,000 men, about twice the size of Britain’s own, who swore their oath not to the crown but to the company.
It was the first company state, a corporation that was also a sovereign, answerable only to its shareholders. For more than a century it looted and ruled much of India. After winning the right to collect Bengal’s taxes in 1765, it raised taxes during the famine of 1769 and 1770, which killed at least a million (and perhaps as many as 10 million).
No AI company has a quarter of a million people under arms (judging by the way drones dominate the battlefield in Ukraine, they might not need them). But OpenAI has partnered with the defense firm Anduril Industries to counter enemy drones, and signed a Pentagon contract worth up to $200 million last year. Anthropic and the Department of Defense are at war over its refusal to let its models power fully autonomous weapons, even as its systems were reportedly used in the capture of Venezuelan President Nicolás Maduro.
And the AI companies haven’t been shy about converting their economic heft into political influence. The Democratic primary for New York’s 12th Congressional District was in some ways a “proxy war” for the AI giants, with various firms or executives of firms — OpenAI, Palantir, Anthropic — supporting different candidates. In the end, AI billionaires spent more than $29 million on the race.
The East India company would have nodded in recognition. Game respects game. In the 1770s, nearly a quarter of the House of Commons held its stock or drew its pay. Its power was broken eventually; after the Indian rebellion of 1857, Britain nationalized the company and took its lands and army. But it had ruled much of India for a century by then, and Indians had paid for it in blood and treasure.
New York’s primary was the AI power struggle breaking out into the open. On the surface, the contest was a draw for the industry: Yes, the winner (who was also endorsed by former Mayor Michael Bloomberg, the founder and owner of Bloomberg LP) was a co-sponsor in the state legislature of the AI safety bill the industry opposed. At the same time, every politician in America now knows that the industry is willing to spend millions against a candidate it doesn’t like. You don’t have to win every fight to exert your power. Fear can do its work for you.
Which brings us to the last thing that makes power so different from wealth — and back to that table in France. Power is zero sum. Everyone can become wealthier. Not everyone can become more powerful. The AI titans represent themselves and their investors. If they have the power to sit as equals at a G7 meeting, then the people who represent everyone else have less power.
Right now, the AI CEOs are at the table by the invitation of world leaders who see them as too important to leave out. But the invitation could one day run the other way. If that’s not the future we want, the time to limit the industry’s power is now. Carbon-based intelligence’s quest for power can be just as dangerous as its silicon progeny’s.
This column reflects the personal views of the author and does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
Gautam Mukunda writes about corporate management and innovation. He teaches leadership at the Yale School of Management and is the author of “Indispensable: When Leaders Really Matter.”
Disclaimer: This report is auto generated from the Bloomberg news service. ThePrint holds no responsibility for its content.