The first cohort to grow up with generative AI as a workplace staple is now bearing the brunt of its economic consequences. New data from the Stanford Digital Economy Lab, released on 19 August 2026, shows that Americans aged 22 to 25 working in roles with high AI exposure have seen their employment numbers fall by 11 percent since ChatGPT launched in November 2022.
That decline stands in stark contrast to low-exposure occupations, which grew by 10 percent over the same stretch. The resulting 19-percentage-point gap has widened from 15 points in summer 2025, signalling that the divergence is accelerating rather than stabilising.
The Stanford analysis draws on payroll records from ADP covering between 3.5 million and 5 million workers. While overall US employment rose 6 percent in recent years, AI-exposed roles lagged at 4 percent growth. A joint scenario modelling exercise by IAB, BIBB and GWS suggests the employment contraction for the youngest workers in these fields could reach as high as 16 percent.
Employers Signal Pay Cuts for New Hires
Wage expectations are shifting just as sharply. A June 2026 Ifo survey of more than 3,000 German companies found that 48.3 percent anticipate lower starting salaries for workers without a university degree and under five years of experience. For more seasoned staff in that bracket, 40.4 percent of firms still expect reduced pay.
The outlook darkens further for young university graduates with less than five years on the job: 50.8 percent of employers project falling wages for this group, while only 16.3 percent foresee increases. Experienced academics fare somewhat better — 26 percent of companies expect pay rises for them, against 37.5 percent predicting cuts.
Sectoral breakdowns reveal where the pressure is most acute. In services, 53.3 percent of firms forecast lower entry-level wages. Retail follows at 47.9 percent, manufacturing at 46.5 percent, and construction at 39 percent.
A Paradox at the Heart of AI Adoption
One of the more counterintuitive findings comes from OpenAI’s own research: entry-level workers are the heaviest users of AI tools like ChatGPT. Yet a Harvard study tracking job advertisements found that postings for roles considered automatable have dropped 13 percent since late 2022, while listings for AI-complementary positions jumped 20 percent. The picture is one of young workers embracing the very technology that is eroding their employment prospects.
As automation reshapes roles across industries, employers face new obligations to keep their teams safe and compliant. A free Health & Safety Toolkit provides ready-to-use risk assessments and checklists covering key UK regulations like the Health & Safety at Work Act 1974 and COSHH. Download the free Health & Safety Toolkit
Global Ripple Effects
The trend extends well beyond the US and Germany. South Korea’s central bank, the Bank of Korea, reported on 19 August 2026 that 295,000 jobs held by people aged 15 to 29 had disappeared, with 268,000 of those — 94 percent — concentrated in AI-intensive industries.
Global youth unemployment stood at 12.4 percent in 2025, according to the International Labour Organization. In Germany, the situation for highly qualified young people has deteriorated markedly: DZ Bank data shows the number of unemployed university graduates under 30 more than doubled between 2022 and 2025.
The cumulative picture suggests a fundamental recalibration of the entry-level labour market — one where the tools designed to boost productivity are simultaneously deflating the value of the workers most eager to use them.
