Global Survey: Most People Expect AI to Destroy More Jobs Than It Creates

Picture a factory worker in Seoul, a call-center employee in Warsaw, and a college graduate in Sydney. They speak different languages and live under different governments. But according to new research from Pew Research Center, they likely share the same quiet worry: that artificial intelligence is coming for their livelihood, not to make it easier.

That worry isn't fringe anxiety. It's the majority view. In 34 of the 37 countries Pew surveyed between February and May of this year, more people believe AI will lead to fewer jobs than believe it will create more. The findings, drawn from interviews with more than 42,000 adults across 36 countries plus additional U.S. samples, paint a picture of a world bracing for economic disruption before most people have even had time to fully understand the technology reshaping their lives.

This isn't a fringe sentiment confined to a handful of anxious nations. It's a pattern stretching from wealthy democracies to middle-income economies, though the intensity varies sharply depending on where you live, how old you are, and how much you already know about AI.

Who Feels the Squeeze Most

Wealth, it turns out, doesn't buy comfort here. Quite the opposite. Countries with higher GDP per capita tend to report the deepest fears about AI-driven job loss. In Australia, South Korea, and the United States, roughly seven-in-ten adults or more expect the technology to eliminate jobs over the next two decades. Think of it like standing closer to a wildfire: the people whose economies are most saturated with AI tools and automation are the ones who can already feel the heat.

In the U.S., the only country where Pew has tracked this question over time, concern about AI-related job loss has climbed 7 percentage points in just two years. That's not a blip. It's a trend line moving in one direction.

Age matters too, and not in the way many might expect. In countries like Canada, France, Singapore, Sweden, India, Indonesia, and Malaysia, younger adults (ages 18 to 34) are more worried about AI-driven job loss than their elders. Meanwhile, older adults in several nations are more likely to simply say they're unsure how AI will affect employment, a sign that awareness itself, not just optimism, shapes how people process this technology.

The generational shift is especially stark in the U.S. Concern about job loss has grown fastest among Americans under 35. Since 2024, the share of young Americans who say they're more concerned than excited about AI jumped from 40% to 55%. For context, that's a swing large enough to erase what used to be a generational gap: five years ago, older Americans were the more worried group. Now the young have caught up, and in some ways overtaken them.

Gender differences also show up in the data. In 11 mostly high-income countries, women are more likely than men to predict AI will shrink the job market. In France, the gap is 10 percentage points. It's a reminder that economic anxiety isn't evenly distributed, even within the same national borders.

Then there's the inequality question, which cuts even deeper than jobs alone. Across the countries surveyed, people are more likely to say AI will widen the gap between rich and poor than to say it will narrow it. Almost no one thinks AI will shrink inequality: in 29 of the nations polled, fewer than one-in-ten adults believe that outcome is likely.

Political ideology plays a role here too. In eight of 28 countries where Pew measured ideological leanings, people on the left were more likely than those on the right to say AI will widen the wealth gap. The starkest example is the United States, where liberals are 31 percentage points more likely than conservatives to hold that view, one of the largest ideological splits in the entire study. Age compounds this in America specifically: adults under 35 are far more likely than those 50 and older to predict AI will deepen inequality, 56% compared to 38%.

Yet fear isn't the whole story. Across 37 countries, a median of 41% of adults say they feel equally concerned and excited about AI's growing presence in daily life, a nuance that gets lost when headlines focus only on dread. Only in Israel do primarily excited respondents outnumber primarily concerned ones. South Korea offers an interesting wrinkle: despite ranking among the most pessimistic nations on job loss and inequality, most South Koreans (61%) describe themselves as equally concerned and excited about AI overall, and 21% say they're mainly excited. Worry about economic disruption, in other words, doesn't always translate into blanket dread about the technology itself.

Awareness seems to soften some of that dread, too. In many countries, people who say they've heard or read a lot about AI report feeling more optimistic than those with little exposure to the topic. Germany illustrates this clearly: among Germans highly familiar with AI, just 26% say they're primarily concerned, compared with 41% among those with minimal awareness. Familiarity, it seems, doesn't always breed contempt. Sometimes it breeds a bit of calm.

But that calm is not universal, and it's not permanent. Concern about AI has climbed meaningfully over the past year in Sweden (up 9 points), Poland (up 8), the Netherlands (up 7), Hungary (up 6), and Australia (up 4). And in several of these countries, younger adults are driving much of that increase, closing gaps with older generations that existed just a year or two ago.

Why It Matters

These numbers matter because public sentiment shapes political will, and political will shapes policy. If large swaths of the global population believe AI will cost them their livelihoods and widen inequality, that belief will eventually show up at the ballot box, in labor negotiations, and in how governments choose to regulate the technology. Ignoring that anxiety, or dismissing it as unfounded, risks a backlash that could stall the very innovations policymakers hope to encourage. Understanding who feels most exposed, and why, is the first step toward building safety nets, retraining programs, and honest public conversations that meet people where their fears actually are, rather than where experts assume they should be.