OXFORD—The United States is both the home of the AI revolution and among the places where people like the technology the least. According to Ipsos’s 2026 AI Monitor, 85 per cent of respondents in China, 80 per cent in India, and 77 per cent in Indonesia think that AI products and services offer more benefits than drawbacks, compared to only 38 per cent in the US, with similarly low shares in Britain, France, and The Netherlands.

 

Such attitudes are likely to shape the technology’s adoption and long-run implications for growth. In early 19th-century Britain, the Luddites famously smashed the mechanical looms that they saw threatening their livelihoods. Parliament responded by making machine-breaking a capital offense. The army was deployed to quash the riots, and Britain went on to become the workshop of the world.

 

By contrast, in China, craft guilds were able to block automation in cotton, silk and other industries until the turn of the 20th century. While China’s failure to industrialsze alongside Europe had many causes, economists Klaus Desmet, Avner Greif, and Stephen Parente identify the power of organised craftsmen to resist labour-saving machinery as a key factor.

 

Since any society that is reluctant to deploy a technology will not capture its gains, it is worth understanding why the AI revolution is being welcomed in some places and resisted in others. This is not a simple East-West divide, considering that only half of Japanese respondents see more benefits than drawbacks from AI.

 

Nor is it the result of Silicon Valley-style doomerism. In the latest Expert Survey on Progress in AI, published in September, the median AI researcher assigned a 10 per cent chance to the technology causing human extinction or a comparably severe loss of control. But researchers educated in Asia felt that such risks were higher than those educated in North America or Europe. The experts closest to the technology in the world’s most AI-optimistic region are, if anything, more alarmed than their Western peers. AI optimism does not reflect a failure to consider what could go wrong.

 

What about perceived economic risks? If richer societies are more invested in the status quo, they may stand to lose more from technological disruption. But better-educated, high-income Americans are more enthusiastic about AI, not less. According to a 2025 Quinnipiac poll, 60 per cent of American households earning more than $200,000 expect AI to do more good than harm in their daily lives, compared with 26 per cent of those earning less than $50,000. But this pattern is reversed in a cross-country comparison: several middle-income societies are among the most optimistic, while Americans, Germans and Canadians are among the wariest.

 

Might this reflect recent growth figures, rather than income levels? While hundreds of millions of people across Asia have seen their incomes rise, Americans, Europeans, and Japanese have spent the same decades wringing their hands over stagnant wages, hollowed-out industrial towns, and fears that their children will be worse off than they were. Harvard University’s Stefanie Stantcheva and her co-authors find that the more economic growth people experienced while coming of age, the less likely they are to see the world in zero-sum terms. If one group’s gain is another’s loss, a technology that reshuffles work is a threat.

 

There is much truth to this. Nevertheless, there is less enthusiasm about AI in fast-growing economies like Poland and Ireland than in Spain and Italy, which have grown far less; and Argentina and South Africa are among the most optimistic, despite a generation of economic stagnation.

 

The most plausible explanation is that AI means very different things depending on what a society already has. In rich economies, it competes with expensive human expertise; but when introduced elsewhere, it supplies expertise that was never available. A patient in Lyon, France, who consults an AI assistant is comparing it with a physician she could have seen instead, while a patient in the Sahel might compare it to a distant clinic or, most likely, nothing at all.

 

The labour-market consequences will be very different, too. In Western hubs like London and New York, the chief worry is that generative AI will put pressure on the incomes of lawyers and accountants, for whom the technology makes hard-won expertise less scarce. But a rice farmer in Bihar (India) who has never seen an agronomist will have an altogether different outlook. What is not to like about an AI system that monitors weather forecasts, soil conditions, crop diseases, and market prices to tell you what to plant, when to irrigate, and whether a yellowing leaf is a blight? AI does not replace anyone on the farm. It gives the farmer an expert they never had.

 

Perhaps more importantly, by narrowing the productivity gap between novices and experts, as numerous studies show, AI may drive a new wave of offshoring to precisely the countries whose citizens feel most enthusiastic about it. GPS technology once made memorising every street unnecessary, allowing anyone with a driver’s license to work for Uber, and forcing incumbent professional drivers to take a pay cut of around 10 per cent.

 

Other professional services now face their own Uber moment, but with one critical difference: Unlike a taxi ride, knowledge work can be imported. The opportunities for wage arbitrage could be vast. A Big Four accountant earns $76,000 in New York and $3,000 in Cairo. With machine translation eroding the language barrier that once confined offshoring to the English-speaking world, the work can go almost anywhere.

 

We are already seeing this happen. The law firm Clifford Chance has cut one-tenth of its London business services staff, citing AI, and shifted work to lower-cost offices in India. Meanwhile, the Big Four have cut UK graduate intakes by up to one-third since 2023, while increasing their hiring in India, Malaysia, and the Philippines. The losers, in other words, are the young graduates in the West who are already most pessimistic about AI. Like the Luddites, they are not irrational enemies of progress. They can just see that they do not stand to benefit.

 

But the Industrial Revolution and China’s own experience are reminders that such sentiments can have long-lasting consequences. While Britain resolved its Luddite problem with troops and nooses, that will not be the West’s approach to AI resistance. To benefit from the productivity boom that AI promises, the countries that built it must ensure that those it threatens share in the gains.

 

Carl Benedikt Frey, associate professor of AI & Work at the Oxford Internet Institute and director of the Future of Work Programme at the Oxford Martin School, is the author, most recently, of “How Progress Ends: Technology, Innovation, and the Fate of Nations” (Princeton University Press, 2025).