Entry-Level Pay in AI-Exposed Fields Has Fallen 13% — and Computer Science Graduates Are Feeling It Most
Published on 09/27/2026 at 22:30 | Editorial boerse-global.de
New research from the US Census Bureau paints a sobering picture for young people entering the workforce in fields most exposed to artificial intelligence. Since ChatGPT arrived on the scene in late 2022, graduates in highly AI-exposed disciplines have been starting their careers earning roughly 13% less than previous cohorts did.
The working paper drew on about 6.7 million bachelor's degrees awarded between 2016 and 2024. Computer science, mathematics and statistics emerged as the disciplines hit hardest.
Employment figures tell a parallel story: the employment rate after graduation in these subjects came in about 5 percentage points lower. The Census Bureau was careful to note, however, that a robust causal link to AI technology has not yet been statistically established.
Where the Missing Pay Goes
About half of the observed wage gap traces back to graduates gravitating toward lower-paying industries. The penalty also fades with time — after a few years of work experience, the 13% shortfall narrows to roughly 5%.
Similar patterns are surfacing in major tech hubs. A report from the New York Center for an Urban Future found that entry-level tech job postings in New York have dropped 49% since 2022 — the steepest decline of any occupational field the researchers examined.
In 2023, for the first time since data collection began in 2010, New York City counted more computer science graduates than available entry-level positions. Citywide, roles with high AI exposure fell 29%, while low-exposure positions grew 21%.
Big technology employers have reshaped their hiring playbooks too. According to SignalFire data, new graduate hiring at 15 leading tech companies fell by more than half by 2025 compared with 2019. Entry-level workers now make up 7% of all Big Tech hires, down from 15% before the pandemic.
An IZA working paper adds another data point: since generative AI became widely available, junior software developer roles have shrunk 14% to 15% relative to senior positions — a sign that employers are raising the experience bar.
Germany's Mixed Signals
The picture in Germany is more nuanced. Stepstone data shows advertised entry-level jobs in the first quarter of 2025 still running 45% below the five-year average.
The first half of 2026 brought some stabilization, with a 14% increase over the same period a year earlier — yet the level remained 18% below the long-term average. On average, graduates needed 40 applications to land a single interview invitation.
Pay in specialized IT fields, meanwhile, has held up. Stepstone puts the gross median salary across the IT occupational group at €66,750.
Experienced specialists earn considerably more, according to Robert Half: IT security architects command €81,250, machine learning engineers €87,000, and AI developers around €92,000. Software architects top the list at up to €109,500.
What Comes Next
With US college graduate unemployment at 7.3% in summer 2026, economists at UCLA warn that cohorts from 2027 onward could face stronger displacement effects. Amazon chief Andy Jassy has said that greater use of generative AI will likely reduce staffing needs in administrative and management functions in the years ahead.
At the same time, the World Economic Forum projects a net gain of 78 million jobs worldwide by 2030 through technological change. In a Bitkom survey of 850 companies, 42% of respondents expected AI to create additional demand for specialists, while about 27% anticipated job cuts.
The emphasis is shifting toward upskilling — but uptake lags. Over the past 12 months, nearly 44% of employees spent no working time at all on AI training, often citing a lack of available courses or insufficient information about them.
