"IA Empires": Niall Ferguson and Eric Schmidt see AI as an imperial technology by nature

In Foreign Affairs, historian Niall Ferguson, former CEO of Google Eric Schmidt and researcher Lisa Klaassen compare China-American rivalry on AI with empire history.

Foreign Affairs magazine publishes a lengthy essay co-authored by Niall Ferguson, a historian specializing in empires, Eric Schmidt, former CEO of Google, and researcher Lisa Klaassen. Their thesis: the term "AI superpowers" does not capture what Washington and Beijing are building. By controlling the big language models, chips, infrastructure and capital, the two countries would build veritable empires, whose authority extends over knowledge itself and makes other states dependent, even without alliances or constraints.

"Imperial" technology in two ways

The authors distinguish two dimensions. On the one hand, the United States (military primacy, network of alliances, reserve currency) and China (industrial domination, the world's largest navy, control over infrastructure in Africa, Asia and Latin America) already had imperial traits, which AI is extending. On the other hand, AI would be imperial in itself: those who master cutting-edge models obtain leverage over science, production and culture, in the manner of money for Habsburg Spain or oil for the "American century". What changes, according to them, is the speed: the cycle of expansion, overextension, exhaustion, which once took a century, now takes place on the scale of a few years.

The lesson of the Dreadnought

History reminds us that empires rarely fall under the blows of a rival: they exhaust themselves defending their possessions, financially (the United Kingdom after 1945), institutionally (the Qing dynasty at the end of the nineteenth century) or technologically. The authors cite the battleship Dreadnought of 1906, which made the existing British fleet obsolete and forced London to almost double its naval spending between 1900 and 1914, to the detriment of its new social promises. Hence their conclusion: the decisive question between Beijing and Washington may not be so much who will reach artificial general intelligence first as who will be able to finance the race to the end.