Skip to content
Slow Variables

Ask the data

Answers come from the same store as the site; every number is checked against the record it cites.

Adaptation · Labour & consumers

Employment in AI-exposed occupations, year on year

consistent with normal80/95 confidence, good evidence, some ambiguitygrade Ccoincident

1What this measures

Year-on-year employment change in the most AI-exposed occupation quintile versus the least exposed, from the Stanford DEL and ADP Research Canaries Dashboard (payroll data). The band applies to the most-exposed series.

Why it matters. The paper predicts no broad AI-attributable employment decline for years; this is the highest-frequency payroll test of that claim.

Proxy types
deployment, welfare
Unit
share
Cadence
quarterly
Valve
adoption to adaptation

2How we track this

  • series adp_research.us_high_exposure.employment_yoy.m
  • series adp_research.us_low_exposure.employment_yoy.m
  • series fortune.us_high_exposure.employment_yoy.m
  • source ADP Research, Canaries Dashboard releases · default tier 6 · ADP Research; short quotation
  • source Fortune · default tier 5 · Fortune Media; short quotation
Normal band
≥ -1.0% and ≤ 1.0%
Fast band
≤ -3.0%
Falsifying
—

Normal = within one percent of flat, ordinary occupational churn; fast = a 3% or worse decline in exposed occupations, which is what a broad AI-attributable displacement looks like in payroll data. Between is `emerging`.

Applied to adp_research.us_high_exposure.employment_yoy.m.

3Tracker interpretation

Exposed occupations are flat while unexposed grow slowly; the signal is in the young, not the aggregate.

4Evidence

Latest point
-0.2%as of 2026-06-30
us_high_exposure
Value the bands apply to
-0.2%as of 2026-06-30

3 observations. Hollow points are disputed (see counterevidence). Every point links to its observation.

5Status and reasoning

consistent with normalsince 2026-09-10 · evaluate

Evaluator: -0.002 (ADP Research, Canaries Dashboard releases, 2026-06-30) is inside the consistent band (lo=-0.01 hi=0.01). Auto-reason; band rationale: Normal = within one percent of flat, ordinary occupational churn; fast = a 3% or worse decline in exposed occupations, which is what a broad AI-attributable displacement looks like in payroll data. Between is `emerging`.

6Timeline notes

  • 2026-06-30 us_high_exposure · -0.2%as of 2026-06-30

7Counterevidence

What cuts against this reading

Exposure is a model-derived measure; ADP payroll coverage skews to larger firms; the divergence began before ChatGPT on some cuts (NY Fed; Iscenko and Millet).

8Update history

  1. 2026-09-10unmeasured to consistent with normalconf — → 80 · evaluate

    Evaluator: -0.002 (ADP Research, Canaries Dashboard releases, 2026-06-30) is inside the consistent band (lo=-0.01 hi=0.01). Auto-reason; band rationale: Normal = within one percent of flat, ordinary occupational churn; fast = a 3% or worse decline in exposed occupations, which is what a broad AI-attributable displacement looks like in payroll data. Between is `emerging`.

    obs:cbb77dc6

9Confidence

80 / 95 — good evidence, some ambiguity

Confidence is independent of status: 90–95 multiple strong independent sources; 70–89 good evidence, some ambiguity; 50–69 mixed or hard to operationalise; below 50 limited or vague.

10Related