Pew surveyed 42,151 adults in 36 countries from February 8 to May 13, 2026, and added two U.S. waves on the American Trends Panel. The report landed on September 17 under the title Globally, More People Expect AI to Cause Job Loss Than Growth, written by Laura Silver, Richard Wike, Moira Fagan, and Jordan Lippert (Silver et al., Pew Research Center, 17 Sep. 2026).
A median of 55 percent of adults in 18 high-income countries say artificial intelligence will lead to fewer jobs in their country over the next 20 years. The middle-income median is 36 percent. Middle-income publics also leave a larger share blank: 34 percent say they are not sure, against 22 percent in high-income countries. Australia, South Korea, and the United States sit near the top of the job-loss column, around seven in ten or higher.
Those country medians are not a population-weighted world figure. They also are not a model of employment. They are what adults say they expect. In the United States, the only country Pew has asked this job question before, the share expecting fewer jobs has risen 7 points in two years.
People who have read more still expect fewer jobs
High-income countries report more awareness. A median of 50 percent say they have heard or read a lot about artificial intelligence, against 27 percent in middle-income countries. In about half the countries, people with high awareness are more likely to expect job losses. The same group is also more likely to say they are equally concerned and excited, or more optimistic overall.
That pairing is the useful part of the file. Adults who follow the subject can hold two thoughts at once: the tools work, and the labor market will still rearrange. A change program that only proves the tool is faster will talk past the second thought.
Young adults already treat the first job as unstable
In every country Pew broke out by age, adults 18 to 34 are more likely than people 50 and older to have heard a lot about the technology. They are also more likely to hold mixed views. In several countries they are more likely to expect job loss. The Philippines is the sharp split: 36 percent of adults under 35 expect fewer jobs, against 11 percent of adults 50 and older. Older adults there are more often unsure, not confident that jobs will grow.
The United States trend is the one hiring managers should keep on the desk. Among Americans 18 to 34, the share who are more concerned than excited rose from 40 percent in 2024 to 55 percent in 2026. That group now matches older adults on primary concern. Five years ago the older group sat higher.
Entry work is drafting, first-pass analysis, and synthesis. Those are the tasks current tools already take. A junior hire who can see that will not cheer a rollout that only advertises the removal of busywork. They need a path that still leads to seniority: review, context, customer judgment, and the coordination of systems other people cannot see.
South Korea expects the shock and still backs the build
South Korea ranks high on expected job loss and on expected inequality. It ranks low on people who say they are primarily concerned. Sixty-one percent say they are equally concerned and excited. Twenty-one percent say they are primarily excited. Israel is the only country in the set where primary excitement beats primary concern.
A workforce can accept a hard labor story and still support adoption if the story is competitiveness, not a personal convenience pitch. A global rollout that copies a U.S. slide deck into Seoul, or copies a Seoul slide into a middle-income market where a third of adults say they are not sure, will miss the room.
Inequality worries fewer people than job loss
The high-income median saying artificial intelligence will widen the gap between rich and poor is 35 percent. The middle-income median is 22 percent. No country has more than one in four saying the gap will shrink. In 29 countries that share is in single digits. In the United States, liberals are 31 points more likely than conservatives to say inequality will rise, one of the largest ideology gaps Pew recorded. Job-loss concern travels more evenly across that split.
Women are more likely than men to expect job losses in 11 countries, mostly high-income ones. France shows a 10-point gap. Pew reports the pattern and does not claim a cause.
What you can do with an unforecastable career
I do not know which titles survive. Neither does the survey. A four-year degree still matters as a first block of method. It does not cover the next 30 years of tools. Budget time and money for a second and third stretch of study while you are employed, not after a layoff letter.
If you run hiring, stop telling new graduates that the tool will clear the dull work so they can do the interesting work. Name the next rung: who reviews model output, who owns the customer when the draft is wrong, who keeps the record when two systems disagree. If that rung is missing, the 55 percent figure among young American adults will show up as quiet attrition, not as a comment on a town hall slide.
If you run a company that sells into more than one income group of countries, split the enablement. High-awareness markets need governance and redeployment language. Lower-awareness markets still need a plain demonstration that the tool finishes a task they already do. South Korea-style markets can hear a competitiveness case without being told to relax.
What happens next
Watch whether the U.S. youth concern number stays at 55 percent in the next Pew wave, or keeps rising. Watch whether companies publish internal mobility numbers next to seat counts for Copilot, Claude, and Gemini. Watch whether middle-income “not sure” shares fall because people see the tools at work, or because they start to expect the same job-loss story the high-income countries already tell.
A related U.S. survey on this site ran a different instrument: 10,980 adults, fed into paths for output and cognitive jobs through 2030. The median path added about 8 percent to GDP and still cut some knowledge work (Americans Expect AI to Add 8 Percent to GDP by 2030). Pew is the public mood around that kind of arithmetic.
Sources
Silver, Laura, et al. “Globally, More People Expect AI to Cause Job Loss Than Growth.” Pew Research Center, 17 Sep. 2026, www.pewresearch.org/global/2026/09/17/globally-more-people-expect-ai-to-cause-job-loss-than-growth/.
Silver, Laura, et al. “What Do Young People Around the World Think About AI?” Pew Research Center, 17 Sep. 2026, www.pewresearch.org/global/2026/09/17/what-do-young-people-around-the-world-think-about-ai/.
Silver, Laura, et al. “Concerns About AI Are Especially Widespread in High-Income Countries.” Pew Research Center, 17 Sep. 2026, www.pewresearch.org/global/2026/09/17/concerns-about-ai-are-especially-widespread-in-high-income-countries/.
Pew Research Center. “Young Adults in the U.S. Are Increasingly Wary of AI, Concerned It Will Take Jobs.” 18 Aug. 2026, www.pewresearch.org/short-reads/2026/08/18/young-adults-in-the-us-are-increasingly-wary-of-ai-concerned-it-will-take-jobs/.
