fellowintel.
LATAM + GLOBAL MACRO
Back to the intelligence desk
Technology / Narratives

AI exposure is not a forecast of disappearing jobs.

A 2026 ILO methodological brief explains why capability-based exposure measures cannot establish actual employment losses.

Research edition. Reporting period: 2026 methodological review of AI exposure indicators. Dates, observations, and forecasts are distinguished. This is a curated snapshot.

The narrative under examination

Editorial framing to examine: an AI-exposed job is necessarily a job that will disappear. This is FellowIntel’s analytical proposition, not an attributed quotation.

01 / THE EVIDENCE

What the sources establish

  • The ILO says exposure indicators measure tasks AI could affect.
  • Results vary with task definitions and measurement assumptions.
  • These measures omit adoption constraints and economic feasibility.
02 / OUR INTERPRETATION

How to read the bigger picture

The claim that an exposed occupation will disappear skips several steps. Deployment, organizational choices, regulation, training, and demand determine whether a capability changes tasks, hours, employment, or wages.

03 / LIMITS & UNCERTAINTY

What we should keep in view

  • The source does not provide a new numerical job-loss forecast.
  • Static task descriptions can become outdated.
  • Exposure alone cannot identify a causal effect on employment.
04 / SIGNALS TO FOLLOW

What would change the picture?

  • Seek observed employment and wage changes, not exposure alone.
  • Distinguish augmentation from substitution in actual workplaces.
05 / OPEN KNOWLEDGE

Read, question, share.

This original editorial analysis is available under CC BY 4.0. Credit FellowIntel, link to the briefing and license, and identify changes. Source materials retain their own rights.

See our methodology