Research check · 31 October 2026
Norway’s exposed jobs did not show a clear employment drop
The October AI Labor Market Index compares Norwegian private-sector employment in the most and least AI-exposed fifths of occupations. Through July 2026, the gap since the pre-ChatGPT baseline was +1.5 percentage points for workers aged 21–60, with a standard error of 2.7 points. That is too imprecise to establish a meaningful difference.
Young workers need a closer look
For ages 21–30, the same comparison was −0.8 points, with a standard error of 4.0. Measured from February 2025 instead, their gap was −2.9 points, with a standard error of 2.0. The later comparison could point to weaker entry, but it is not conclusive. The Norwegian age band is wider than the 22–25 group in the U.S. Stanford–ADP payroll study.
What the index counts
Norwegian employers report paid jobs monthly. The index covers private-sector wage earners aged 21–60 and groups 397 occupations by how much of their task mix language models could speed up. It compares three-month averages and provides seasonal, population and exposure-measure alternatives. The authors’ June paper ends in February 2026; the October dashboard updates the same research, not an independent replication.
What it cannot prove
The exposure score does not record which Norwegian employers used AI. Other demand changes, demographics and occupation shifts could move the lines. The broad Norwegian result therefore does not refute a narrower U.S. hiring warning, and the younger Norwegian estimate does not establish AI-caused displacement. A country comparison needs the same age band, employment measure and adoption data before it can test the mechanism.