The promise of artificial intelligence is often framed as universal abundance—a future where productivity surges and everyone benefits. But history suggests a different outcome: a splintering of prosperity, where the gains concentrate among capital and high-skilled workers while the majority see stagnant wages. Adam Smith, writing in The Wealth of Nations over 200 years ago, foresaw this tension. His anecdote of a playful boy who automated a steam valve by tying a string to its handle captures both the genius and the dark side of automation: the boy gained freedom but likely lost his job. Smith described this as a machine improvement, yet the boy’s perspective was one of obsolescence. This duality—efficiency gains versus human displacement—remains central to understanding AI’s impact today.
Smith also warned about the dehumanizing effects of extreme division of labour in the fifth book of The Wealth of Nations, noting that workers could become "as stupid and ignorant as it is possible for a human creature to become." Productivity and human flourishing are not automatically aligned. Furthermore, in The Theory of Moral Sentiments, he argued that the pursuit of wealth is fundamentally about social recognition, not material necessity. A rich person’s stomach can only hold one meal, but the desire for status is limitless. These two insights—that production efficiency and distributional fairness are separate, and that status competition drives behaviour—are crucial for predicting the AI era.
The historical record of the first Industrial Revolution is sobering. Economic historian Robert Allen coined the term "Engels’ Pause" to describe the period from 1790 to 1840 in Britain, where per capita GDP grew by 46% but real wages for workers increased only 12%. The surplus flowed to factory owners, not laborers. The story of Robert Blincoe, a child textile worker born in 1792, illustrates this brutality. He was apprenticed at age seven to a cotton mill where he worked 14-16 hour days, cleaning machinery while it was running, often getting his hands caught. His memoir, published in 1832, helped spur the 1833 Factory Act, the first meaningful restriction on child labour in Britain. The gap between productivity and wages wasn’t closed by market forces—it was closed by political intervention.
Today, AI is following a similar pattern. A 2024 paper by MIT economist Daron Acemoglu, "The Simple Macroeconomics of AI," estimates that AI will contribute only 0.93%-1.16% to US GDP over the next decade, with labour’s share declining because current AI deployment focuses on automation rather than creating new tasks. The consulting firm McKinsey projects that up to 30% of work activities could be automated by 2030, disproportionately affecting lower-wage jobs in retail, food service, and manufacturing. Meanwhile, the wealth of tech billionaires has surged—Forbes reports that the top five tech fortunes added over $200 billion in 2023 alone. The root problem is that automation rewards the owners of capital and human-like intelligence, not the providers of routine labour.
A contrasting perspective comes from economist Erik Brynjolfsson, who argues that AI can also augment human capabilities, creating new roles in fields like healthcare, education, and creative industries. However, this requires deliberate investments in retraining and institutional redesign, which are currently lacking. The "Engels’ Pause" of the Industrial Revolution lasted half a century before legislation caught up. For AI, we haven’t yet begun the political and social adaptation needed to ensure broad-based prosperity. Smith defined national wealth not as money, but as "the necessaries and conveniences of life." If we only measure GDP growth without tracking who gains access to those necessities, we repeat the mistake of the 1790s. The question is not whether AI will produce more, but whether the mechanisms of distribution—taxation, social safety nets, education, and worker representation—will evolve fast enough. Without that evolution, the coming prosperity will remain splintered, enriching a few while leaving many behind.