The current issue of Foreign Affairs focuses on AI and on global fragmentation, two regular SIGnal themes. The lineup of authors on AI is especially formidable, including Mohammed El-Erian, Ruchir Sharma and Niall Ferguson as well as people more commonly associated with the topic (Eric Schmidt, James Manyika, Alex Imas). Global fragmentation is handled by Henry Farrell and two co-authors, and Farrell has consistently been among the most original of geopolitical thinkers in the past decade. His article and the AI trio are worth reading together, because fragmentation and AI are themselves best understood together.
As several authors point out, AI might have been developed in some modestly cooperative way. Microsoft did, after all, have a serious AI lab in China until recently. But the logic of de-globalization, which predates ChatGPT and the rest, brought the return of a sovereignty, and self-sufficiency, discourse that quickly turned into insistence on protecting national telecoms networks. This was a key policy point for the first Trump administration. Under Joe Biden, digital innovation as such became subject to sovereignty requirements, as it already had been in China. Digital-innovation protectionism had both security and economic drivers. They fused. Farrell et al. capture the result in the excellent phrase “the economic security state.”
Schmidt, Ferguson and Lisa Klaasen graft ideas about empire onto AI reality. They see the US and China as at the head of two “empires of AI,” arguing that “AI is by its nature an imperial technology.” The US and China are alone at the top: “governments across the globe must pick a side.” The authors try out several possible endings to this story of clashing empires. The most persuasive two are a debt crisis and war. The debt crisis comes from the fact of soaring government debt in both China and the US combined with their shared dependence on AI-related industry for much of their economic expansion over the past five years. If AI is getting bubbly, as all the authors in the issue more or less agree, then a recession is all but inevitable as the governments concerned have lost the resources to cushion a clearing-out of AI competitors. “AI,” the authors conclude, “is the most capital-intensive imperial project in history, undertaken by two superpowers that are already weighed down by debt.”
The war — again, this is a thread through the issue — would be over Taiwan, whose semiconductor industry has become so critical, above all in Taiwan itself, whose economy is otherwise unremarkable.
It’s striking how much these assumptions about AI, US-China competition and a Taiwan war have settled into the proudly unhysterical world of Foreign Affairs. Together they form a consensus that will have serious influence if the Republican party performs poorly in the US midterm elections.
The authors are all pretty pessimistic, some extremely so. Henry Farrell and his co-authors note that it “took the Cuban missile crisis and the threat of nuclear annihilation” to force state actors to develop a means of reducing tensions. Schmidt, Ferguson and Klaassen can’t find much hope for a “benign outcome in a world increasingly run by powerful and relentlessly self-improving machines.” It is quite possible that such deep pessimism is a symptom of the current disenfranchisement of the powerful, highly educated and highly experienced stratum of American society that the authors inhabit. They simply don’t influence the Trump administration. This is reflected in how their articles don’t have US policy recommendations, which have been essential to Foreign Affairs articles in earlier eras. When people accustomed to influence lose it, they can become apocalyptic. But if power shifts after Tuesday, the authors’ slice of society should regain influence, which will then engender some much-needed optimism.
That’s the note that El-Erian, Imas and Manyika sound, although they limit themselves to the US. But of course the US is the best place to start.