AI Revolution · Explainer · Updated Oct 2026
AI and jobs: what the data actually shows (so far)
By James Brown · Published · Updated
Not mass unemployment, but not nothing. The early evidence points at entry-level hiring.
The loudest claims about AI and work ("it'll take every job" or "it changes nothing") are mostly vibes. Here's what researchers have measured so far.
The aggregate picture: no big wave yet
A Stanford SIEPR policy brief found little evidence that AI is causing significant job losses so far. Since 2022, unemployment for the most AI-exposed fifth of workers rose 0.77 percentage points, slightly less than the 0.85-point rise for the least exposed. That points to a broadly softening labor market, not AI-specific displacement.
The pressure point: the first rung of the ladder
- An IZA discussion paper (June 2026) using Lightcast job-posting data found a 14–15% relative decline in junior versus senior software developer vacancies after ChatGPT's release, with employers asking for more experience in the same job titles.
- A study using the Census Bureau's Quarterly Workforce Indicators found early-career (22–24) employment in the most AI-exposed industry-state cells fell about 12% over 10 quarters after ChatGPT's release, though hiring largely recovered by early 2025.
- Researchers at LSE's Centre for Economic Performance caution that remote work explains much of the decline in junior hiring. Once work-from-home exposure is controlled for, the measured GenAI effect shrinks sharply.
The other side
PwC's 2026 AI Jobs Barometer, built on more than a billion job ads, reports that headcount growth at the most AI-exposed companies is outpacing the least exposed, and that skills in AI-exposed jobs are changing more than twice as fast. Its read: companies are using AI for growth, not just cost cutting.
What to do with this
If you're early in your career, the bar for "junior" is rising. Show judgment, communication, and evidence that you can ship work, ideally in public. A portfolio site is the cheapest résumé upgrade there is.
Sources
- Stanford SIEPR, What is really happening to jobs?
- Westby, Modestino & Cheng, Generative AI and the Redefinition of Entry-Level Software Work (IZA DP 18723, Jun 2026)
- Tucker, You're (not) hired: AI and early career hiring (Apr 2026)
- LSE CEP, The broken ladder: AI, remote work, and early-career hiring
- PwC, 2026 Global AI Jobs Barometer
More AI explainers See the market snapshot →
FAQ
Is AI causing mass unemployment?
Current evidence says no. A Stanford SIEPR brief found unemployment among the most AI-exposed workers rose slightly less since 2022 than among the least exposed.
Which workers are most affected so far?
Early-career workers in AI-exposed fields such as software development and customer service, mainly through fewer junior job openings.
Is remote work part of the story?
Yes. LSE researchers find remote-work exposure explains much of the decline in junior hiring once both factors are considered together.