Economists surveyed by Indeed expect AI to affect the salaries of graduates, but not those of other workers

123 panellists responded between 8 and 16 September. They believe the risk of displacement is concentrated amongst higher education graduates, and anticipate almost no pressure on others.

Indeed's Hiring Lab published its quarterly labour market forecasting survey on 24 September, conducted with Pulsenomics among 123 economists and experts surveyed between 8 and 16 September. The exercise is interesting less for its forecasts than for the dividing line it reveals.

The responses are expressed in the form of diffusion indices, where 50 indicates the absence of a trend. On the risk of AI displacing jobs, the index reaches 55.6 for workers with a higher education degree, compared to 48.1 for non-graduates — i.e. no perceived pressure for the latter. The panel also expects more downward pressure on the median real wages of graduates, with the corresponding index rising from 47.3 to 42.4 in one quarter.

On productivity, opinions remain measured: 47% expect a moderate improvement, 27% a substantial gain and only 3% a transformative effect. US unemployment forecasts remain flat, at 4.15% in September 2026, 4.23% at the end of the year and 4.32% a year later. AI, in this table, redistributes wages before destroying jobs.