KI als Jobkiller für Berufseinsteiger? Entwarnung mit Vorbehalt / AI as a job killer for labour-market entrants? All-clear with caveats
A · Strong evidence IAB researchers (Arntz, Dinelli, Schlenker) analyse monthly social-security records for about 22.3 million full-time workers and control for telework and recession exposure, with method stated.
Summary based on the full text·Original: about 8 minutes to read
In short
An IAB-Forum article (13 October 2026) by Melanie Arntz, Michele Dinelli and Oliver Schlenker of Germany's Institute for Employment Research. German administrative data show no general employment slump in AI-exposed occupations; overall they grew faster than less exposed ones. For workers aged 20-29 the association with exposure is negative and growing but not statistically significant, and much smaller than the effect of the weak economy. A significant negative link appears only in manufacturing and non-academic occupations.
Key numbers
- ~22.3 millionfull-time workers aged 20-64 covered monthly, January 2021 to December 2025
- -3.8% vs -2.6%change in full-time employment of 20-29 year-olds since 2022, more vs less AI-exposed occupations
- ~1%employment in the two most exposed occupation groups in 2025 above 2022, vs a fall of over 3% in the three less exposed
- 1.8%weaker 2025 employment for 20-29 year-olds per standard deviation of AI exposure, not significant
- 4.2%fall in young employment in occupations in economically weak sectors, 2022 to 2025
As reported by the source.
English·Web page·opens iab-forum.de
What it reports
- 2022 to 2025: full-time employment of 20-29 year-olds fell about 3%; 30-39 about 1.3%; 40-64 under 1%.
- The two most exposed occupation groups were about 1% above 2022 in 2025; the three less exposed about 3% below. For workers 30+ the exposed groups grew about 2%.
- For 20-29 year-olds: -3.8% in highly exposed vs -2.6% in less exposed occupations, a growth gap of -1.2 points; for 30-39 the gap was nearly +6 points and for 40-64 about +5.
- With telework suitability and recession exposure controlled, one standard deviation more AI exposure went with 1.8% weaker 2025 employment for the young (not significant) and 1.4% stronger for those 40+.
- Recession exposure had a strong and growing negative association; young employment in weak-sector occupations fell 4.2%. The authors call it more than twice the AI association.
- Significant negative association for young workers only in manufacturing and non-academic occupations; none in services or academic occupations.
How they got it
Monthly Federal Employment Agency records, January 2021 to December 2025, for full-time workers subject to social insurance (excluding civil servants, self-employed, apprentices and others). Eloundou et al. (2024) exposure mapped to German occupation groups (KldB 2010), telework suitability from Alipour et al. (2023), and recession exposure from mass layoffs and closures in 2022-24. Poisson regressions by age group, occupation and industry, quarterly, with fixed effects.
Limits
- Descriptive; the authors say results only hint that AI may have partly contributed.
- Exposure measures potential use, not whether employers adopted AI.
- Full-time workers only; apprentices and marginal jobs, important for entrants, are excluded.
- Data end in 2025; the authors ask for continued monitoring.
Why it matters for companies
Read alongside the Banque de France finding for France, this shows the youth effect is not uniform across countries and that the business cycle can dominate. For workforce planning in Germany, the warning sign is in non-academic and manufacturing roles, not graduate entry. Attributing weak junior hiring to AI without controlling for demand risks the wrong response.
Key passages in translation
Quoted in English
Ein allgemeiner Beschäftigungseinbruch in KI-exponierten Berufen bleibt aus.
There is no general employment slump in AI-exposed occupations.
Quoted in English
Die Ergebnisse deuten darauf hin, dass vor allem die Wirtschaftsschwäche der letzten Jahre zu einer sinkenden Beschäftigung von jungen Beschäftigten beigetragen hat, während die KI-Exposition unterm Strich eine eher untergeordnete Rolle gespielt haben dürfte.
The results suggest that mainly the economic weakness of recent years contributed to falling employment among young workers, while AI exposure probably played a rather minor role overall.
Quoted in English
Entgegen der öffentlichen Diskussion um die Einstiegsprobleme junger Akademiker*innen lenken die Analysen den Blick daher eher auf nichtakademisch geprägte Berufe und das verarbeitende Gewerbe.
Contrary to the public debate about young graduates' entry problems, the analysis points instead to non-academic occupations and manufacturing.
Summary based on the full text. Read the original for the full argument and data.Read the original on iab-forum.de
- Tags
- Labour market