- Economics
How do worker beliefs about layoff risk due to generative artificial intelligence compare to the realized labor market impact of the technology, and what shapes those beliefs? We field a largescale, multi-wave survey measuring workers’ perceptions of AI-related job loss risk and AI adoption in the workplace, which we estimate at 30–40% of U.S. workers through the first half of 2026. The average U.S. worker, regardless of whether they use the technology at work, believes there is a 20% chance of losing their job to generative AI within two years, which is comparable to workers’ realized risk of job loss from any cause over a similar horizon. Using a continuoustreatment difference-in-differences design, we find that these fears are not borne out, as job postings and layoffs in more exposed occupations show no statistically significant response to the diffusion of generative AI. Instead, fear rises with usage intensity (more hours per week) and is concentrated among those who deploy AI for core occupational functions such as programming rather than for drafting tasks such as writing. These results suggest that workers learn about AI’s capacity to substitute for their labor by using it and form beliefs about displacement independent of any realized job loss.
