AI enters US election polling: fear over jobs is growing, though no figures were published

By Zendoric · September 18, 2026.

Scott Clement's "The Pollwatcher" column in The Washington Post (September 17, 2026) is devoted mainly to a Washington Post-Schar School poll on the Michigan Senate race between Democrat Abdul El-Sayed and Republican Mike Rogers. In passing, it mentions that two new surveys find a growing share of Americans who fear AI and technology will cost them their jobs, and it cites a third poll on the decline in confidence in the measles vaccine since 2020. The text, as accessed, includes no figures, methodology or polling firm for those two AI surveys: it breaks off just before developing them, with a "But first..." that is never resolved in the available content.

Broadly speaking, unease about AI's impact on employment has been entrenched in the US public debate for more than a year, fueled by waves of layoffs in technology, media and services that companies themselves have linked, with varying degrees of honesty, to automation. That this concern now appears folded into the routine political and public-health polling cycle of an outfit like Washington Post-Schar School, alongside the Michigan race and vaccine confidence, is telling in itself: it has stopped being a niche tech topic and become a public-opinion variable that campaign strategists watch as closely as inflation or healthcare.

It is worth separating two levels that the debate frequently blurs. One is declared fear -how many Americans say they fear losing their job to AI-, which measures social anxiety and has immediate political consequences: it feeds protest voting, demands for regulation and distrust of Big Tech, regardless of whether the risk materializes on that scale. The other level is the real impact, sector by sector, which at Zendoric we have spent months documenting: uneven, concentrated first in administrative and back-office tasks, and more respectful of expert judgment, human relationships and on-site work. A poll detecting more fear does not mean displacement is accelerating at the same pace; it means public perception, fed by layoff headlines, is running ahead of the granular evidence on which jobs disappear and which merely change content.

Honesty requires acknowledging that this fear is not unfounded in the short term: there will be a hard transition, with concrete losers in concrete occupations, and playing it down would be as poor an analysis as amplifying it without nuance. But it is worth keeping in mind that this kind of poll will keep recurring as a thermometer of an anxiety that will foreseeably ease only when the abundance AI can generate -more resources, more time, better health- shows up in the material lives of those who today answer a questionnaire with fear. Until then, every poll published without figures of its own is, paradoxically, a data point: the topic already sells headlines on its own, even if the rigor with which it is covered still lags behind the interest it arouses.