China Embraces AI While America Frets

Start with a chart, and sometimes the entire argument is already there.

The Economist's editors gathered recently to discuss a data visualization that mapped public sentiment on artificial intelligence across two axes: excitement on one, nervousness on the other. The picture it produced was stark. Chinese respondents clustered toward enthusiasm, with low nervousness readings. American respondents sat at the opposite end — markedly more anxious, markedly less enthused. The gap wasn't marginal. It was the kind of divergence that suggests two populations aren't just disagreeing about a technology; they're operating from fundamentally different assumptions about what technology does to a society.

The question the editors kept circling was whether that gap reflects genuine structural differences — or whether it's simply a matter of timing.

The History Behind the Mood

The optimism case for China isn't naive, and The Economist's editors resist treating it that way. As editor Corbin put it in the discussion, the Chinese population "has become very adept to technological change, and actually in the past 20 or 30 years technological upgrading has brought massive gains for China. They've only known improving standards of living with technological change."

That's a real observation about a real history. The last three decades in China have been defined by industrialization, urbanization, and export-led growth that pulled hundreds of millions of people out of poverty. Technology — manufacturing technology, platform technology, logistics technology — was the vehicle. From where many Chinese workers sit, the machine has been, on balance, on their side.

Americans carry a different memory. Deindustrialization, the hollowing out of manufacturing towns, stagnant wages alongside soaring productivity — the American relationship with technological change over the same period has been far more ambivalent. Not universally negative, but contested in ways that have real political weight.

So the divergence on the chart isn't a mystery. It's two populations reading the same headline — "AI is coming for jobs" — through very different recent histories.

Shenzhen, On Two Tracks

The broad optimism, though, starts to fracture the closer you get to specific workers in specific jobs. The Economist's Sarah, reporting from Shenzhen — a city that has functioned as a laboratory for technological transformation for four decades — found a society that feels, in her telling, like it's running on two parallel tracks.

Tech workers she spoke to were nervous. Their roles are expanding, their responsibilities multiplying, and the implicit promise that technical skills equal job security is starting to look shakier. There's also what Sarah described as "the curse of 35" — a persistent pattern of age discrimination in Chinese tech hiring that makes the anxiety sharper. In a sector where layoffs are constant and restructuring is the operating mode, being on the wrong side of 35 can effectively end a career trajectory. AI hasn't created that problem, but it doesn't obviously fix it either.

Then there's the other track. Sarah spoke to gig workers — food delivery riders, parcel couriers — the people who the editors noted could be next in line for automation if delivery drones and robotic logistics scale further. Their response, she found, was largely pragmatic: "This is normal. Technology is always going to develop. We'll figure things out."

That pragmatism is worth sitting with. It could reflect a genuine cultural orientation toward adaptation. It could reflect a rational calculation that worry is unproductive when the outcome isn't in your hands. It might also reflect something more constrained — a sense that there are limited channels for expressing structural discontent, so one develops a different relationship with acceptance.

Probably some combination of all three. The editors don't fully resolve this, and they're right not to.

The Political Asymmetry

Here is where the conversation gets genuinely interesting, and where the gap between the two countries becomes more than a matter of attitude surveys.

One of the editors made an observation that deserves to be taken seriously: "In some sense the Chinese political system can be a little bit more able to not accommodate — the second you have a kind of data center uprising, as you do in parts of the US, that affects a primary election, that instantly filters through, then everyone gets very anti-data centers."

The U.S. system — fractious, responsive to local pressure, prone to policy whiplash — is in some ways poorly suited to managing the kind of economy-wide transformation that AI represents. Disruption aggregates into political pressure, which produces reactive policy, which can be captured by incumbents as much as by those actually displaced.

China's system can absorb discontent differently. Beijing can signal priorities, direct investment, and suppress or redirect grievance before it crystallizes into opposition. That's a real form of capacity — but it's also a real form of risk. The same low tolerance for dissent that keeps AI backlash from becoming a political crisis could make it far more explosive if it ever does reach that point. The editors draw a direct analogy to zero-COVID: a policy maintained through authoritarian suppression of dissent that eventually collapsed anyway, at enormous cost.

"You're having a big AI backlash right now," one editor noted. "If any of that moves towards China and collides with the political system that has a very low tolerance for dissent — as we saw during zero-COVID — that could be quite explosive too."

Beijing's Early Moves

Beijing, at least, appears to be reading the same risk. The government isn't waiting for a crisis to begin thinking about labor market disruption. According to China's five-year plan targets on AI's employment impact — as reported by The Next Web — Beijing has built in mechanisms to track how AI is reshaping job creation and destruction across key sectors and cities, with early-warning systems intended to get ahead of mass layoffs rather than respond to them.

Sarah's reporting confirms that government advisors are actively working through the tradeoffs. The dilemma Beijing faces is a familiar one for any government trying to manage technological transition: how do you reassure workers without signaling to companies that they'll be penalized for efficiency? How do you protect people from disruption while staying competitive in a global race where your rivals aren't protecting people from anything?

So far, the approach is mostly surveillance — in the technical sense. Watch closely. Build the data infrastructure to understand what's actually happening. Be ready to cushion impacts before they become crises.

Whether watching closely is sufficient to the scale of what's coming is another question.

The Retraining Dodge

Every serious discussion of AI and labor displacement eventually arrives at the same three-word answer: retraining, education, reskilling. The Economist's editors get there too, to their credit with some skepticism. As one put it: "Everyone sort of comes back to: we have to have retraining, we have to have constant education, we have to make sure that schooling's right. And of course those things are sensible, but whether they match the scale of the task, I really doubt."

That doubt is well-placed. Retraining programs have a modest track record at the best of times. They work reasonably well for workers who are younger, more educated, and geographically mobile — which is to say, the workers who were probably going to be fine anyway. For the 45-year-old courier in Shenzhen, or the 52-year-old data-entry worker in Ohio, "go retrain" is often less a solution than a way of restating the problem in more optimistic language.

This isn't an argument against education or retraining. It's an argument that the policy conversation tends to default to supply-side solutions — improve the workers — when the question of who captures the productivity gains from automation is the harder, less comfortable one to answer.

Both China and the United States are, so far, mostly avoiding that harder question. China is building monitoring infrastructure. The U.S. is arguing about data centers in primary elections. Neither country has found a convincing answer to what an economy looks like when a meaningful share of tasks can be done more cheaply by machines, and what that means for the people whose livelihoods those tasks represented.

The chart that opened The Economist's discussion showed two populations in very different emotional places. What it didn't show — what no chart yet captures — is where both of them end up when the automation wave hits sectors that the optimists weren't expecting.

---

By Raj Mehta, Global Markets & International Finance Reporter