AI Turbo in Operation: Between Job Anxiety and Efficiency Frenzy
(Bild: DC Studio/Shutterstock.com)
A new Weizenbaum study shows great dynamism in AI adoption by companies. But whether employees go along often depends on the involvement of works councils.
The integration of Artificial Intelligence [1] into the German economy has moved beyond the theoretical stage. What was recently considered a future vision has arrived in operational reality in 2025. A study by the Berlin Weizenbaum Institute shows that reservations are diminishing and the hunger for productivity gains is growing. Within just one year, the proportion of companies using AI in regular operations has increased from 50 to 62 percent. If companies currently in pilot projects for the introduction of the technology are included, the rate rises to 74 percent.
According to the study, algorithms are already part of the team in IT, administration, and marketing, among others according to the study [2]. Penetration is high in areas such as IT security and programming, where an average of 55 percent of companies already use AI in regular operations. Usage is also increasing in sensitive areas such as human resources. Here, 25 percent of companies reported usage (18 percent regular operation, 7 percent pilot phase).
However, behind the technological facade, a dispute over the distribution of potential dividends [3] is raging. Almost 80 percent of companies state that they primarily use AI to increase efficiency. The question is what happens to the time gained. Here, the study paints a picture that at least partially allays the widespread fear of rapid mass job losses [4]. While 40 percent of companies are toying with the idea of replacing staff with AI, the reality is currently different.
Quality Leap Instead of Layoff Wave
The vast majority of over 80 percent currently use the freed-up capacity to improve the quality of their products and services. Almost three-quarters of companies use the time saved to reduce the chronic backlog of overtime and thus noticeably relieve the workload of their employees. Instead of primarily using the algorithm as a job killer, it currently acts more as a buffer in a labor market [5] that is already under high performance pressure.
To assess the scientific validity of the results, it is worth looking at the study design: The panel study is based on a stratified random sample from the years 2024 and 2025, in which 440 companies with more than 50 employees from the manufacturing and service sectors were most recently surveyed.
A special feature is the methodological coupling of perspectives. In over 230 cases, the researchers interviewed management and work council chairpersons to obtain a contrasting picture. By excluding micro-enterprises and concentrating on AI-relevant core industries, the study does not represent the entire economy. However, it provides a valid data basis for companies driving digital transformation in Germany.
Co-determination as a Safety Net
Overall, the climate in the companies remains divided. The acceptance of new technology does not fall from the sky but is the result of tough negotiations and transparent communication. The scientists reveal a correlation here: According to them, the success of AI implementation stands or falls with the practiced co-determination [6]. Where management actively involves the works council from the outset, employees report significantly less often about stressful work intensification due to the new systems. Currently, this happens in about 53 percent of the cases studied.
In such cooperatively managed companies, AI is understood as a tool that supports people rather than controlling or pressuring them. If this involvement is lacking, mistrust among employees grows rapidly. While managers typically see the effects on work design through rose-tinted glasses, employee representatives fear a gradual loss of scope for action and increasing disempowerment. This different perception shows how important dialogue between hierarchies is in order not to force technological transformation against employees.
Martin Krzywdzinski from the Weizenbaum Institute sees the results as confirmation of the European digitalization model. The expert involved in the study warns against misunderstanding co-determination as an annoying brake on innovation. While consultation processes with the works council may occasionally slow down the rollout of new systems, this loss of time is a valuable investment in the social sustainability of the company. Ultimately, the analysis indicates that the AI turbo can boost productivity without necessarily worsening working conditions. Prerequisite: The power balance in the company is maintained, and the human remains the focus.
[7](nie [8])
Don't miss any news – follow us on Facebook [9], LinkedIn [10] or Mastodon [11].
This article was originally published in German [12]. It was translated with technical assistance and editorially reviewed before publication.
URL dieses Artikels:
https://www.heise.de/-11228857
Links in diesem Artikel:
[1] https://www.heise.de/thema/Kuenstliche-Intelligenz
[2] https://www.weizenbaum-institut.de/media/Publikationen/Weizenbaum_Discussion_Paper/Weizenbaum_Discussion_Paper_53.pdf
[3] https://www.heise.de/news/Gutachten-KI-fuehrt-deutsche-Wirtschaft-zurueck-auf-den-Produktivitaetspfad-10320259.html?from-en=1
[4] https://www.heise.de/tp/article/KI-und-GPT-Tschuess-ihr-Nutzlosen-8949538.html
[5] https://www.heise.de/news/Bundesregierung-ChatGPT-Co-sind-bislang-keine-Job-Killer-8974599.html?from-en=1
[6] https://www.heise.de/news/Digitalisierung-KI-Bundesrat-will-mehr-Mitbestimmungsrechte-fuer-Betriebsraete-10484817.html?from-en=1
[7] https://www.heise.de/newsletter/anmeldung.html?id=ki-update&wt_mc=intern.red.ho.ho_nl_ki.ho.markenbanner.markenbanner
[8] mailto:nie@heise.de
[9] https://www.facebook.com/heiseonlineEnglish
[10] https://www.linkedin.com/company/104691972
[11] https://social.heise.de/@heiseonlineenglish
[12] https://www.heise.de/news/KI-Turbo-im-Betrieb-Zwischen-Job-Angst-und-Effizienzrausch-11228806.html
Copyright © 2026 Heise Medien