Modern digital technologies can replace some human activities while simultaneously creating new ones. As a result, they change the demand for certain skills in the labor market. The rapid development of generative artificial intelligence (AI)—especially through applications such as ChatGPT—has significantly intensified the discussion about the impact of AI on the labor market, both in acadenua and in public debate and policy.
Key questions in academic research include:
- Is the demand for AI skills also growing on the German labour market?
- Does the increasing demand for AI skills mean that other skills - among low, medium and highly qualified workers - are less in demand?
The 10th ai:conomics Policy Brief outlines how a reliable data foundation is being established within the ongoing research project to enable informed statements about the impact of AI on the demand and supply of skills. It presents the current status of the dataset development, which links online job adverts with German social security register data. Additionally, it highlights which research gaps this innovative data basis can help close in the future and what new insights the ai:conomics team has already gained.
Previous Approaches, Challenges, and the Starting Point of the ai:conomics Project
Some recent studies have measured the impact of AI by using data from online job adverts (Bessen et al, 2023; Gonschor and Storm, 2023; Babina et al, 2024, 2023; Alekseeva et al, 2021; Acemoglu et al, 2022), as these provide a near real-time, forward-looking and largely objective data source for assessing labor demand. However, the challenge with this approach is that in many countries, linking such data with individual labor market data is difficult.
Some notable studies from Europe have addressed these challenges by taking advantage of the fact that public employment agencies in some European countries support companies in publishing job advertisements as part of their employer services (Damelang et al., 2024; Mueller et al., 2024). This allowed researchers to link information from these job advertisements with labour market data from social security registers, as both social security registers and job advertisements are integrated into the administrative processes of the same public service providers. However, this approach also presents challenges: companies are not obligated to use the employer services of public employment agencies. Additionally, jobs requiring a high level of qualification are often not captured in these postings—an important limitation, as research shows that AI particularly affects highly skilled jobs.
This is where ai:conomics comes in: a team of researchers from IAB Nuremberg, ROA Maastricht, FAU Nuremberg, IWH Halle, and Stanford University combined private online job postings from the provider Lightcast with register data from the German social security system comprehensively for the first time. The advantage of using online job postings compared to data provided by employment agencies is that Lightcast aggregates data from numerous websites. This allows for a more detailed overview of the structure of job openings in the German labor market and enables more precise insights into the impact of AI on different skill levels and industries.
