Stanford: AI Job Losses Hit Young Workers Hardest
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Stanford University analysis of four years of US employment data found workers aged 22-25 in AI-exposed sectors saw a 2.7% employment drop since ChatGPT became widespread, rising to 12.8% in the most exposed fields — finance, software and creative industries (with caveats that interest rate effects may explain part of the signal).
- Nobel prize-winning economists have publicly warned the world "must act now" to ensure AI produces rising living standards rather than large-scale job displacement, while London businesses report struggling to find needed skills as AI disrupts the jobs market.
- Token usage at the world's top AI-deploying companies has surged into the trillions — sometimes quadrillions — over recent months, almost entirely for "agentic" task automation, with bills so large that several firms have begun rationing access to the most advanced models.
- OECD analysis of job postings shows the UK was hit notably hard in highly exposed sectors (telemarketing, legal services) at a time when interest rates were stable or being cut, predating last year's National Insurance rise, with its service-sector concentration leaving it exposed on most international measures.
- The latest generation of LLMs can complete complex software tasks that take humans several hours — including auditing cryptocurrency contracts and streamlining their own models — with self-development capability expected within roughly a year, and the same pattern now emerging at an earlier stage in financial analysis, entry-level legal work and creative jobs.
- Many companies, including Western ones, are pivoting to much cheaper AI derived from Chinese models offered freely, introducing a new cost dynamic that the article flags as a key trend to watch for the economics of automation.
- London businesses warned they were struggling to find the skills they need as AI disrupts the jobs market, even as headlines focus on job loss rather than the hiring squeeze created by shifting skill demands.
Why it matters: For employers, the cost math isn't settled: agentic AI at the frontier has become so expensive that companies are rationing it, suggesting the promised labor savings may not materialize at current pricing — especially as cheaper Chinese-derived models reshape the unit economics. For 22-25-year-olds entering software, finance and creative fields, Stanford's 12.8% employment drop in the most exposed sectors is the first hard data point showing the disruption is already landing on a specific cohort, not a future hypothetical.
