ChatGPT's launch raised job-loss worries in five European countries, a survey study finds

The authors used an Unexpected Event During Survey design, which treats the launch as a natural experiment during fieldwork. They matched respondents statistically and ran regressions to compare socioeconomic groups. Concern rose most among people with low and middling income and education, and among those poorly exposed to AI.

The authors say the timing gives a rare view of reactions before informational framings consolidate and before widespread adoption. They conclude that the socioeconomic consequences of AI are perceived as immediately salient. The measured outcome is worry about job loss, not actual job losses.

The study addresses a gap the authors identify: despite the sudden rise in AI's public salience after ChatGPT's launch, they say no empirical investigation had assessed its effect on citizens' primary concerns.

To isolate the launch's effect, the team relied on a survey that happened to be in the field when ChatGPT was released. They compared respondents before and after the launch using statistical matching, then used regression analyses to examine differences between socioeconomic groups.

The authors report that the effect appeared in advanced industrial democracies and was concentrated among people with low and middling income and education, as well as those poorly exposed to AI. They describe the timing as capturing reactions before widespread adoption.