Will AI take our jobs? The answer may not be so simple. Currently, there is a clear pattern in the labour market: younger people are finding it increasingly hard to get jobs in AI-related fields, while employment among older workers is rising. This is supported by new research from Örebro University and Ratio.
The new study indicates that changes in the labour market are not mainly about jobs disappearing, but about who gets them. The overall decline in job advertisements is primarily linked to a weaker economy, rather than AI. However, behind this trend, a pattern emerges: younger workers are finding it increasingly difficult to enter professions where AI is commonly used.
We're observing a significant drop in employment among young people, which appears to be partly driven by the rise of generative AI. This decline is especially apparent in fields heavily dependent on AI, such as software development and customer service roles," says Magnus Lodefalk, docent and senior lecturer in economics at Örebro University and affiliated researcher at Ratio.
The study is based on Swedish register data covering the entire workforce from 2019 to June 2025. Within the same employer, employment in AI-exposed occupations among people aged 22 to 25 declined by 6.5 per cent compared to less-exposed occupations after the launch of ChatGPT. Meanwhile, employment for individuals over 50 in the same occupations increased by 1.5 per cent.
The pattern is evident: the younger the worker, the more pronounced the negative impact.
Our results suggest that generative AI does not primarily reduce the number of jobs but rather affects who gets them. In AI-exposed roles, we see a relative decline among young people but not among older adults. This suggests that AI primarily impacts the process of entering the labour market," says Magnus Lodefalk.
Researchers point to the necessity for new pathways into the labour market for young people, such as AI-integrated vocational training programmes and internships.
The study is a collaboration between Magnus Lodefalk and Lydia Löthman (Örebro University), Erik Engberg (Örebro University and Ratio), and Michael Koch (Aarhus University and Kiel Institut für Weltwirtschaft). Research is conducted within AI-Econ Lab and the WASP-HS-funded research cluster AISCAF.
Read the study here (PDF).