Adaptation · Labour & consumers
Entry-level employment shortfall in AI-exposed occupations
How far employment of 22–25-year-olds in the two most AI-exposed quintiles has fallen short of keeping pace with the three least exposed, per the Canaries paper (Aug 2026 revision).
Reason through thisThe reading
Widening (15% to 19% in a year) and concentrated exactly where the paper says AI bites first. Real, but not yet the broad decline the falsification rule needs.
2 readings drawn from stanford_del.us_entry_level_exposed.employment_gap.pt, one of 3 series listed under How it is measured; 4 observations in the evidence.
Source: Stanford Digital Economy Lab. Brynjolfsson, Collis, Eggers, Kazinnik & Nguyen, 'What is Generative AI Worth?' (2026).
Why it matters
If AI substitutes for the codified knowledge new graduates bring, the first sign is hiring, not separations. This is that sign.
What cuts against it
Timing partly predates ChatGPT; hiring freezes in tech and finance explain part of it; the measure is relative, so exposed-occupation youth could be flat while others boom.
How it is measured
- series
stanford_del.us_entry_level_exposed.employment_gap.pt - series
stanford_del.us_entry_level_exposed.employment_change_since_nov2022.pt - series
stanford_del.us_entry_level_unexposed.employment_change_since_nov2022.pt - source Stanford Digital Economy Lab · default tier 6 · Working paper
- proxy deployment, welfare · cadence quarterly · valve adoption to adaptation
Normal = a shortfall under 10 points, inside the range youth employment swings over a cycle; fast = 25 points or more. The current reading sits between, so it is `emerging`.
Applied to stanford_del.us_entry_level_exposed.employment_gap.pt.
Latest readings
Why this status, and its history
Canaries paper, August 2026 revision: the kept-pace shortfall for 22-25-year-olds in the two most-exposed quintiles is 19% as of June 2026, up from 15% at the July 2025 vintage; in levels, -11% vs +10% for the least exposed since November 2022. Between the normal band (<=10%) and the fast band (>=25%). Initial seed.
- 2026-09-10unmeasured to emergingconf — → 70 · initial seed (model-drafted)
Canaries paper, August 2026 revision: the kept-pace shortfall for 22-25-year-olds in the two most-exposed quintiles is 19% as of June 2026, up from 15% at the July 2025 vintage; in levels, -11% vs +10% for the least exposed since November 2022. Between the normal band (<=10%) and the fast band (>=25%). Initial seed.
source1
Why this confidence
Rests on 4 observations from Stanford Digital Economy Lab; best evidence tier 6, grade B; current.
Confidence is independent of status: 90–95 multiple strong independent sources; 70–89 good evidence, some ambiguity; 50–69 mixed or hard to operationalise; 0–49 limited or vague evidence.
Predictions that lean on this
- Too early to tellSubstantial case: output rises and some office jobs go by 2030. · Korinek et al. (The Anthropic Institute, WP 2026-02)