jobs worldwide with some GenAI exposure
Exposure is not displacement.
ILO, 2025Evidence Checks
Evidence before prediction
Hiring, adoption, productivity, policy and worker experience, with the measurement limits shown.
of surveyed organisations using AI
Agent deployment remains early.
Stanford AI Index, 2026of global employment in the ILO's highest exposure band
Impacts differ sharply by gender and income.
ILO, 2025Complete evidence checks
Read the method, limits and practical meaning
These guides distinguish measured change from forecasts and confident claims.
AI Job-Loss Predictions: What the Evidence Actually Says
Large numbers usually describe exposed tasks, employer expectations or gross job movement, not a verified count of people who will lose work because of AI.
Are There Really AI-Proof Careers?
No credible occupation is guaranteed immunity, but some work has a stronger mix of demand, human responsibility and difficult-to-automate context.
How Long Does a Job Search Really Take?
There is no honest universal duration. Official unemployment data describe incomplete spells, while your search depends on role, region, route and constraints.
The Hidden Job Market: Useful Strategy or Misleading Myth?
Some opportunities are filled before or without a public advert, but dramatic percentages are usually unsupported and can distract from a balanced search.
Are People More Productive at Home? An Evidence Review
There is no universal home-productivity effect. Results vary with the job, schedule, coordination, measurement and employee choice.
Does the Four-Day Week Work? What Trials Really Found
Trials show promising wellbeing and retention results, but success usually follows work redesign and the evidence is not a universal productivity guarantee.
Observed22 Jul 2026
One in four jobs has some GenAI exposure. That is not the same as one in four jobs disappearing.
The ILO's task-level index points to transformation as the likelier outcome and shows why clerical work deserves immediate attention.
Read the evidence note →Observed21 Jul 2026
AI adoption reached 88% of surveyed organisations. Broad job losses still do not show up in the aggregate data.
Adoption, useful deployment and workforce reduction are different measures. Treat them separately.
Read the evidence note →