AI job loss statistics 2026

The index in numbers

Figures from the AI Job Risk Index, which publishes a score only when at least two distinct research sources cover an occupation. How scoring works.

• 961 occupations hold a published AI displacement risk score, built from 20 distinct research publications.
• The mean published score is 31 out of 100; the median is 29.
• 0 of the 961 scored occupations score 70 or higher, the high risk band. 281 sit between 40 and 69, and 680 fall below 40.
• The most exposed occupation in the index right now is Customer Service Representatives at 68 out of 100. The least exposed is Tapers at 7.
• The most exposed sector by average score is Technology at 55 out of 100 across 38 occupations; the least exposed is Construction & Trades at 13.
• Sources disagree more than most coverage admits: across occupations covered by more than one source, the median gap between the highest and lowest individual estimate is 34 points, and 567 occupations have estimates at least 30 points apart.

What published research projects

The figures most often quoted in coverage, each with its origin. Sources we can point at are linked; the rest are named. These are estimates built on different methods, and they do not all measure the same thing.

• 205,000 US workers were covered by job cuts citing AI or automation as a driver in the first eight months of 2026, per ResumePulse tracking reported by Outsource Accelerator, matching the whole of 2025 in under eight months.
• 49 percent of current customer service jobs are projected by Forrester analysts to disappear by 2030, per reporting on the support industry.
• Workers aged 22 to 25 in the most AI exposed occupations saw a 16 percent relative employment decline, while senior colleagues in the same occupations held steady, per the Stanford payroll study Canaries in the Coal Mine.
• The World Economic Forum's Future of Jobs survey expects employers to shed 92 million roles and create 170 million new ones by 2030, a net gain of 78 million, though not for the same people doing the same tasks.
• Goldman Sachs estimated in 2023 that generative AI could expose work equivalent to around 300 million full time jobs globally. Exposure means tasks touched, not jobs deleted.
• McKinsey's faster scenarios put roughly a third of current work hours in reach of automation by 2030.
• An MIT study of computer vision automation found that at current costs only about a quarter of the wages attached to technically automatable vision tasks would be economical to automate today. Capability is not adoption.

Terms these statistics use

AI exposure

How much of an occupation's work current AI could in principle perform or meaningfully assist. An exposure figure is not a prediction that the jobs will disappear, and published exposure studies split into two families that barely correlate: overlap with what language models do, and automation potential across all tasks.

AI displacement risk score

Our 0 to 100 occupation score: a weighted consensus of published research estimates, calibrated onto one scale, describing the typical version of a role rather than any individual. The full method and its limits are on the data page.

High risk, transitioning, lower risk

The index's three bands: 70 or above, 40 to 69, and below 40. Most occupations sit in the middle band, where parts of the job automate and the job changes shape rather than disappearing.

Citing these figures

Index figures on this page are free to cite with attribution to the AI Job Risk Index and a link to this page or to the ranking. They change as research is ingested, so quote the date shown above with the number. Figures attributed to other organisations should be cited to those organisations, not to us. For the cuts actually announced this year, see our analysis of the 2026 layoff wave; for what the timelines mean for one person, see when will AI take your job.