Künstliche Intelligenz: Potenzielle Effekte für den deutschen Arbeitsmarkt

Abstract

"The ongoing development and integration of artificial intelligence (AI) is increasingly shaping markets and workplaces both globally and within Germany. The present scenario analysis addresses the potential impacts of AI on the development of the German economy and labour market over a 15-year period. For this purpose, an AI scenario was developed and then compared to a reference scenario. In the AI scenario, the necessary infrastructure for AI development and integration must be in place, offering sufficient computing capacity in data centres. Furthermore, there will be different economic-sector-specific penetration paths and efficiency levels. Companies in the respective economic sectors have to invest upfront. Consequently, the impact of AI will vary depending on the economic sector. In some cases, material savings can be achieved. In others, productivity increases or new sales opportunities can be generated. In many economic sectors, a combination of these economic potentials is conceivable. It is evident that other countries are also engaged in the development and integration of AI. For the AI scenario, it is therefore assumed that the overall economic effects abroad will be similar to those in Germany, with the corresponding effects on German import prices. The results show that the implementation of AI can result in a substantial enhancement in added value, particularly over the medium and long term, while maintaining almost equivalent labour input as in the reference scenario. Based on the assumptions made, there is potential for annual economic growth to be 0.8 percentage points higher on average than in the reference scenario. Cumulatively over the next 15 years, 4.5 trillion euros in additional value added could be generated. In the AI scenario, total employment reaches a level similar to that of the reference scenario after 15 years. Nevertheless, there are clear shifts in the labour market. Certain economic sectors will experience an increase in labour demand, while others face a decline in employment. In the initial stages of accelerated AI development and integration, an overall higher labour demand can be expected. The higher labour demand is driven by the need to establish the necessary infrastructure, process and prepare data, and develop models. In the medium term, there will be a slightly lower demand for labour due to increased efficiency. However, towards the end of the projection period, there is potential to compensate for these medium-term employment losses. IAB-Forschungsbericht 23|2025 7 The development of new business models in particular will result in higher demand for labour. Overall, the opposing effects on employment will ultimately be balanced by each other. Hence, the assumptions made in the AI scenario do not exclusively lead to positive or negative effects on the labour market but indicate that there will be shifts in some economic sectors. In IT and information service activities, for example, the number of employed persons is after 15 years around 110,000 higher than in the reference scenario. In contrast, the number of employed persons in business support activities is around 120,000 lower. Nevertheless, a decline in employment numbers should not necessarily be interpreted as a deterioration in the labour market. In the long term, scarce human resources could be deployed more efficiently with the potential to reduce labour shortages in affected occupational groups. At the same time, it is to be expected that an increased use of AI will change the requirements within the workplace. The effect on employment by requirement level indicates that highly skilled jobs are particularly more affected by AI applications than by traditional, non-learning software." (Author's abstract, IAB-Doku) ((en))

Cite article

Zika, G., Hassemer, T., Hummel, M., Krebs, B., Maier, T., Mönnig, A., Schneemann, C., Weber, E. & Zenk, J. (2025): Künstliche Intelligenz: Potenzielle Effekte für den deutschen Arbeitsmarkt. (IAB-Forschungsbericht 23/2025), Nürnberg, 58 p. DOI:10.48720/IAB.FB.2523