For 22-to-25-year-olds in the jobs AI touches most, employment is 19 percent below where it should be. For everyone else it is fine.
The August update of Stanford's 'Canaries in the Coal Mine' study, by Erik Brynjolfsson, Bharat Chandar and Ruyu Chen, follows payroll records through June 2026 and finds no economy-wide displacement — but a widening hole at the entry level. Employment for 22-to-25-year-olds in the most AI-exposed occupations sits about 19 percent below where it would be had it kept pace with same-age workers in less exposed jobs, up from 15 percent a year earlier. In absolute terms that group's employment fell about 11 percent between November 2022 and June 2026, while less exposed categories grew about 10 percent; experienced workers in jobs built on tacit, learned-on-the-floor knowledge gained. The authors are explicit that these are descriptive patterns, not causal proof, and that the gaps shrink once education is accounted for.
Related stories
One state hired a single full-time person for AI in schools. Two years later he has trained a third of its teachers.2026-08-25A university just banned AI in its core course — for the teachers as well as the students, including for grading
2026-08-24Harvard Business School will sell you eight weeks for $699 — and the instructor giving you feedback is a copy of one
2026-08-23Vietnam has written down exactly how a student may and may not use AI — and it is not a ban
2026-08-2226,811 pupils, 30 months: homework marks rose 18 percent and exam results fell 20
2026-08-22