Pew Survey Finds Most People in 37 Countries See AI as a Job Threat
A 42,151-person Pew Research poll across 37 countries shows majorities fear AI will eliminate jobs and widen inequality, with concerns highest in middle-income nations.
A new Pew Research survey of 42,151 people in 37 countries finds that most respondents view artificial intelligence as a threat to employment, with worries about job displacement outpacing optimism about the technology's benefits. The poll, conducted from February 8 to May 13, was completed before a wave of recent public warnings from AI researchers and executives about existential risks. Pew released the findings this week, offering one of the broadest snapshots yet of global public opinion on machine learning.
The survey's timing matters. Respondents answered months before figures like Geoffrey Hinton left Google to speak freely about AI dangers, and before OpenAI's Sam Altman testified to the U.S. Senate. That means the job anxiety Pew measured was not a reaction to those headlines; it was already baked into public attitudes across dozens of economies. The consistency of the concern across countries suggests it stems from lived experience with automation and labor-market shifts, not from a single news cycle.
Income inequality ranked as another major worry. Majorities in many surveyed nations said AI would make the gap between rich and poor worse, according to Pew's breakdown. Middle-income countries tended to register stronger anxiety than wealthier ones, a pattern Pew has observed in earlier technology surveys. The finding complicates the common assumption that AI skepticism is mainly a phenomenon of rich-world knowledge workers facing disruption.
Pew's sample covered a wide range of economies, from advanced industrial nations to emerging markets, giving the results unusual geographic reach. That breadth is a departure from most AI-attitude research, which tends to draw heavily on respondents in the United States and Western Europe. The 37-country scope means the report can speak to how workers in manufacturing hubs, service economies and agricultural regions separately assess the same technology.
What the survey does not settle is whether these fears will harden into policy demands or fade as AI tools prove more assistive than replacing. Pew fielded the questions before generative AI products reached mass adoption in many of the countries surveyed, so the numbers capture a baseline rather than a verdict. Researchers will want to rerun the questions now that tools like ChatGPT are in daily use. The open question is whether exposure to the technology changes minds or confirms the worry.