China, Germany, and AI: It's Not Wealth, It's Expectation

China trusts artificial intelligence more than almost any other country measured. Germany trusts it less by the day. Optimism about AI is a reading of one's own economic position because it reflects how much each population expects to gain from the technology. The easy interpretation holds that Germans are prudent and the Chinese naive. Or that wealth makes some cautious and need makes others optimistic. That reading is comfortable. It's also wrong.

The Ipsos AI Monitor 2026, the fifth edition of the study, surveyed twenty-three thousand five hundred thirty-two adults across thirty-two countries between March and April. Asia leads in optimism, with China at the top. Western Europe and the United States are markedly more cautious. In seventeen of the twenty-eight countries measured in both years, people are now less likely to think AI brings more benefits than drawbacks. Enthusiasm dropped eight points in Germany, eight in Britain, seven in France, and five in the United States.

The version already circulating in columns and panels is appealing because it fits intuition. The more resources a country has, the more it distrusts the technology promising to transform its economy. The wealthy have something to lose. Everyone else doesn't. Caution would then be wisdom accumulated by societies that already lived through an industrial revolution. Concrete experience.

Germany lived through the automation of its manufacturing industry and knows the cost in jobs. France has a social protection apparatus that AI threatens to render obsolete. Britain went through its own version of technological transformation with the deindustrialization of the 1980s. These are societies with consolidated institutions, unions with historical memory, and a certain learned allergy to promises that later leave communities stranded. Their caution is born of experience, not ignorance.

Ipsos itself, in its 2025 edition, identified something more precise than wealth as the variable. The most enthusiastic countries tend to be the ones that most believe AI will benefit their own economy. That's not a correlation with GDP per capita. The correlation lies in the expectation of personal benefit. And that's where the easy argument starts to fall apart. Wealth and expectation of benefit are not the same thing. Confusing them is the central error of nearly this entire conversation.

This distinction matters because it changes who we should be looking at. If the variable were only wealth, we'd expect all rich countries to be equally distrustful. That's not what happens. Singapore is wealthy and ranks among the most enthusiastic. South Korea, a developed economy, tempers its optimism more than Indonesia does. Why? The variable isn't how much money a country has. It's whether its population believes the next wave will benefit them or someone else.

In countries where labor is already organized around consolidated institutions—unions with history, mature pension systems—AI is perceived as a threat to something that exists and functions, however imperfectly. In economies where a huge share of the population works outside that formal architecture, AI threatens nothing they already have. It can instead be the path to something they never had. In The Generosity in the Doorway I develop this same tension from Mexico's vantage point. About 55% of the workforce operates in the informal economy. Any technological promise of income or efficiency arrives in a country where half the population isn't even registered in the systems that would administer that benefit. The question isn't whether AI helps or harms in the abstract. It's who it finds already protected and who it finds exposed.

There's another piece almost no one connects to the first. The Digital News Report 2026 records the lowest trust in news since measurement began in 2015, just 37% globally. People form their opinion of an opaque technology—one not even its creators explain clearly—through a news system they also don't trust. They don't evaluate with clean data. They evaluate through media, influencers, and forwarding chains they already distrust from the start. A double opacity disguised as informed opinion.

Ipsos itself found that 46% of respondents would trust a generative AI less if they knew its answers were influenced by advertisers, with a majority in fifteen of the markets measured. People rightly suspect that a chatbot's answers might be colored by whoever is paying. The same applies to what they hear from the press about how safe or dangerous the technology is. There's money behind both versions—the one selling caution to justify restrictions and the one selling optimism to justify investment. International safety declarations arrived without binding mechanisms attached. That pattern benefits precisely those with the most resources to produce that discourse.

What's missing from the public conversation is recognizing that the narrative about AI isn't a neutral reading of a technology. It's a tool of economic policy. If enthusiasm depends on how much the respondent expects to gain, then convincing a population that AI is a threat or an opportunity has strategic value. Those with the most resources to fund studies, columns, campaigns, and forums are almost always the same ones who profit most from a specific version. It's no coincidence that the letter signed by lab employees calling for tools to slow development came from the same companies racing to reach first the frontier they claim to want to restrain.

There's no automatic virtue in German caution nor automatic naivety in Chinese optimism. Both stances are rational responses to different economic positions. A South Korean worker who sees AI threatening a stable job is correctly reading their exposure. A young person in Indonesia who sees AI as a gateway to global markets previously out of reach is also correctly reading their own lack of exposure to what they fear losing. Neither is evaluating the technology itself. They're evaluating their position relative to it, with the imperfect information they have, through media that, by their own account, they trust less than ever.

The data point that contradicts the initial intuition is this: the country where enthusiasm fell the most between 2025 and 2026 is not the poorest nor the hardest hit by automation. It's Germany. And the country where that optimism holds strongest is not Asia's wealthiest. It's China. If wealth explained the distrust, both cases would be reversed. Where do we stand in this equation?

1. Ipsos AI Monitor 2026 (fifth edition), June 2026

2. Ipsos AI Monitor 2025

3. Digital News Report 2026, Reuters Institute for the Study of Journalism