Why do tech moguls seem so disconnected from the real world? Think, for example, of Anthropic co-founder Dario Amodei saying that half of entry-level jobs could be wiped out by AI in a few years, while also predicting a 95% reduction in cancer mortality and incidence in the same time period. Or Elon Musk burbling on about how AI will make work “optional”.
Why do these guys talk such nonsense? Is it just ignorance of the world? Or is it the arrogance that comes from the reality distortion field that surrounds the unimaginably wealthy; those who inhabit a zone in which nobody dares to disagree with them, so they wind up thinking they are geniuses?
Answer: a bit of both, but the main reason is that they are, overwhelmingly, males who come from narrow Stem (science, technology, engineering, and mathematics) backgrounds and are trapped in a particular mindset that the technology writer Evgeny Morozov called “solutionism”.
This is an ideology that recasts complex social phenomena such as politics, public health, education and law enforcement as “neatly defined problems with definite, computable solutions or as transparent and self-evident processes that can be easily optimised – if only the right algorithms are in place!”
Solutionism has been endemic in Silicon Valley from its earliest days, but the arrival of AI has dramatically turbocharged it. This technology, we are told, will enable us to solve the climate crisis, eliminate cancer, boost productivity, quadruple gross domestic product, transform education and so on. In short, AI will change the world.
Which naturally leads you to wonder what the transformed world envisaged by the tech industry may look like. With impeccable timing, the economist Branko Milanović, an expert well known for his work on income distribution and inequality, has helpfully provided us with an outline sketch of this brave new nirvana envisaged by Silicon Valley.
For information about how The Observer protects your data, read our Privacy Policy
Milanović thinks that untrammelled deployment of AI would create a society that is highly stratified into four social classes. At the bottom of the heap is what he calls the “surplus population”; for example, people whose skills are made obsolete by AI, either permanently or in a way that will require them to spend years retraining. Silicon Valley thinks of these as humans who have little or no industrial value.
If this group reaches between 15% and 20% of the labour force, Milanović thinks, it could become politically explosive, needing to be kept “quiet and docile” through modern bread and circuses (lifetime unemployment benefits, plus 24/7 low-cost streaming entertainment, perhaps?).
Then there will be a “service-sector working class” comprising those whom AI hasn’t yet replaced, living precariously on low-to-moderate wages under constant threat of being pushed down into the surplus population.
Above them is what Milanović calls the “homoploutic” class; those who are simultaneously in the top 10% of capital owners and top 10% of wage earners (currently about 3% of the US population). Think of them as “professionals”; lawyers, surgeons, dentists, corporate executives, readers of the Financial Times and the Economist. He thinks that AI will probably increase the size of this group, since its high labour-plus-capital income enables it to keep buying into the most profitable sectors of the new economy.
And, finally, at the very top of this stratified pyramid, Milanović’s model envisages a new aristocracy of tech lords: an extraordinarily wealthy elite, consisting of thousands of people (not just the 0.001% we have now) who own the technology itself and wield outsized political influence.
The revealing thing about this model is that it produces a logical extrapolation of Silicon Valley’s fantasies about AI. It shows how the vision of AI-driven abundance and leisure that Musk recently outlined to the Economist’s bemused editor-in-chief, Zanny Minton Beddoes, results instead in a chronically unequal and politically unstable society. It’s a dystopia created not by malicious AI, but by how its economic gains are (and are not) shared.
Photograph by Bloomberg via Getty Images