Most People Think AI Will Cut Jobs, Not Create Them

More people in 34 of 37 countries think artificial intelligence will lead to fewer jobs in the next 20 years, not more jobs, according to Pew Research Center survey data published September 17, 2026.

The report, titled "Globally, More People Expect AI to Cause Job Loss Than Growth," is based on interviews with 42,151 adults in 37 countries from February 8 to May 13. Pew Research Center The sample is large for this kind of research. The gap is wide.

In only three countries did views not lean toward job loss. In the rest, expecting loss was the most common answer. The share was highest in Australia at 76%, South Korea also at 76%, and the United States at 71%. The Verge USA Today reported the same main finding.

Respondents were also more likely to say AI will widen the gap between rich and poor. This was true around the world, not only in richer countries where companies already use AI systems the most.

Views on daily life with AI were more mixed. Across countries, a median of 41% said they feel equally concerned and excited about AI in daily life. A median of 37% said they feel mostly concerned, compared with 13% mostly excited.

Age makes a difference. Older people were more likely to say they are concerned about AI in daily life. But concern among younger adults about job loss and inequality rose fast in the past year in Sweden, Poland, Japan, Australia, Brazil and the United States.

Earlier Pew research gives background on who has heard of AI. Pew found a clear link between national income, measured by GDP per capita, and awareness of AI, in research published October 15, 2025. Pew Research Center In a separate survey of 25 countries, a median of 37% said they trust the United States to regulate AI well, while 48% said they do not. In the U.S. itself, half of adults said in a June 2025 survey that more AI in daily life makes them feel more concerned than excited.

The broader context here helps explain the numbers. Public expectations often move faster than what companies have actually adopted. People also group together different changes, like machines doing routine tasks, hiring freezes, moving junior office work elsewhere, software that sets shifts and reviews performance, and pay gains that go mostly to some groups. Think of AI here like office software that learns patterns from large amounts of text and data to draft, answer and decide. The survey does not separate those causes. It shows a general mood about direction, not a firm estimate of size or timing.

In my view, that mood should be taken seriously without being treated as a final answer on the technology. I watched my own two children go from doubting AI helpers to using them daily for school and early jobs, in months rather than years. Familiarity did not remove worries about cheating, loss of skills or entry-level hiring. It made those worries more concrete. For people who build technology, the message is less about how capable the systems are and more about whether people trust them. Systems that affect hiring, scheduling, reviews and access to expert help will be judged on jobs and fairness, as well as accuracy, response speed and safety. Over the long term, I remain hopeful these tools will help small teams and individuals do more, but that will depend on choices that give workers clear leverage, not only employers clear savings.