The Self-Other Gap in Perceived Automation Risk: Evidence from the United States and Canada

This visualization shows a systematic misperception in how workers judge automation risk. Drawing on the 2026 Measuring Employment Sentiments and Social Inequality study, the authors compare paired measures of perceived automation likelihood for self and most others, using nationally representative samples of American and Canadian workers. Approximately three quarters of study participants in both countries rated their own jobs as at low risk of automation in the next few years, yet more than 70 percent believed that most other workers face at least some likelihood of automation. The pattern aligns with pluralistic ignorance: most workers hold one view of their own automation risk while assuming that most others hold a different one. The self-other gap is invariant across occupational categories and across two distinct national contexts, consistent with an informational asymmetry in which beliefs about others’ risk reflect prevailing public narratives about artificial intelligence rather than workers’ direct experience.