Zitatform

Giuntella, Osea, Johannes Konig & Luca Stella (2025): Artificial intelligence and the wellbeing of workers. In: Scientific Reports, Jg. 15, H. 1. DOI:10.1038/s41598-025-98241-3

Abstract

"This study explores the relationship between artificial intelligence (AI) and workers’ well-being and healthusing longitudinal survey data from Germany (2000–2020). Using a measure of occupational exposure to AI, we explore an event study design and a difference-in-differences approach to compare AI-exposed and non-exposed workers. Before AI became widely available, there is no evidence of differential pre­trends in workers’ well-being and health. We findno evidence of a sizeable negative impact of AI on workers’ well-being and mental health. If anything, there is evidence of an improvement in health status and health satisfaction, which may be explained by the decline in job physical intensity. Overall, our results are consistent with the lack of negative effects of AI on the labor markets." (Author's abstract, IAB-Doku) ((en))

Zitatform

Gmyrek, Pawel, Janine Berg, Karol Kamiński, Filip Konopczyński, Agnieszka Ładna, Balint Nafradi, Konrad Rosłaniec & Marek Troszyński (2025): Generative AI and jobs. A refined global index of occupational exposure. (ILO working paper / International Labour Organization 140), Geneva, 72 S. DOI:10.54394/hetp0387

Abstract

"This study updates the ILO’s 2023 Global Index of Occupational Exposure to Generative AI (GenAI), incorporating recent advances in the technology and increasing user familiarity with GenAI tools. Using a representative sample from the 29,753 tasks in the Polish occupational classification system and a survey of 1,640 people employed in each 1-digit ISCO-08 groups, we collect 52,558 data points regarding perceive potential of automation for 2,861 tasks. We then compare this input with a survey and several rounds of Delphi-style discussions among a smaller group of international experts. Based on this process, we create a repository of knowledge about task automation that goes beyond national specificities and use it to develop an AI assistant able to predict scores for tasks in the technical documentation of ISCO-08. Our 2025 scores are presented in a revised framework of four progressively increasing exposure gradients, with a new set of global estimates of employment shares exposed to GenAI. Clerical occupations continue to have the highest exposure levels. Additionally, some strongly digitized occupations have increased exposure, highlighting the expanding abilities of GenAI regarding specialized tasks in professional and technical roles. Globally, one in four workers are in an occupation with some GenAI exposure. 3.3% of global employment falls into the highest exposure category, albeit with significant differences between female (4.7%) and male employment (2.4%). These differences increase with countries’ income (9.6% female vs 3.5% male in Gradient 4in HICs), and so does the overall exposure (11% of total employment in LICs vs 34% in HICs). As most occupations consist of tasks that require human input, transformation of jobs is the most likely impact of GenAI. Linking our refined index with national micro data enables precise projections of such transformations, offering a foundation for social dialogue and targeted policy responses to manage the transition." (Author's abstract, IAB-Doku) ((en))

Zitatform

Gmyrek, Pawel, Christoph Lutz & Gemma Newlands (2025): A technological construction of society: Comparing GPT-4 and human respondents for occupational evaluation in the UK. In: BJIR, Jg. 63, H. 1, S. 180-208. DOI:10.1111/bjir.12840

Abstract

"Despite initial research about the biases and perceptions of large language models (LLMs), we lack evidence on how LLMs evaluate occupations, especially in comparison to human evaluators. In this paper, we present a systematic comparison of occupational evaluations by GPT-4 with those from an in-depth, high-quality and recent human respondents survey in the UK. Covering the full ISCO-08 occupational landscape, with 580 occupations and two distinct metrics (prestige and social value), our findings indicate that GPT-4 and human scores are highly correlated across all ISCO-08 major groups. At the same time, GPT-4 substantially under- or overestimates the occupational prestige and social value of many occupations, particularly for emerging digital and stigmatized or illicit occupations. Our analyses show both the potential and risk of using LLM-generated data for sociological and occupational research. We also discuss the policy implications of our findings for the integration of LLM tools into the world of work." (Author's abstract, IAB-Doku, Published by arrangement with John Wiley & Sons) ((en))

Zitatform

Golboyz, Mark (2025): Governing the Digital Transition: The Moderating Effect of Unemployment Benefits on Technology‐Induced Employment Outcomes. In: Social Inclusion, Jg. 13. DOI:10.17645/si.10114

Abstract

"The digital transition shapes work in numerous ways. For instance, by affecting employment structures. To ensure that the digital transition results in better employment opportunities in terms of socio-economic status, labor markets have to be guided appropriately. The European Pillar of Social Rights can be the political framework to foster access to employment and tackle inequalities that result from the digital transition. Current research primarily examines scenarios of occupational upgrading and employment polarisation. In the empirical literature, there is no consensus on which of these developments prevail. Findings vary between countries and across different study periods. Accordingly, this article provides a theoretical explanation for the conditions under which occupational upgrading and employment polarization become more likely. Further, this article examines how the use of information and communication technology (ICT) capital in the production of goods and services affects the socio-economic status of individuals and, more importantly, whether unemployment benefits moderate this effect. Methodologically, the article uses multilevel maximum likelihood regression models with an empirical focus on 12 European countries and 19 industries. The analysis is based on data from the European Labour Force Survey (EU-LFS), the European Union Level Analysis of Capital, Labour, Energy, Materials, and Service Inputs (EU-KLEMS) research project, and the Comparative Welfare Entitlements Project (CWEP). The results of the article indicate that generous unemployment benefits are associated with occupational upgrading. This implies that educational and vocational labor market policies need to be developed to prevent the under-skilled from being left behind and to enable these groups to benefit from the digital transition. Consequently, it is not only the extent to which work involves routine tasks or the skills of workers that determine how technological change affects employment, but also social rights shape employment through unemployment benefits." (Author's abstract, IAB-Doku) ((en))

Zitatform

Greenan, Nathalie, Dario Guarascio & Jelena Reljic (2025): AI and the labour market: opening the black box. In: Eurasian business review, Jg. 15, H. 4, S. 925-951. DOI:10.1007/s40821-025-00324-8

Abstract

"This work aims at discussing some of the main (open) questions about the labour impact of AI technologies. First, we provide an in-depth literature review focusing on concepts and measurement approaches and distinguishing between up (invention and knowledge creation), mid (technological innovation and development) and downstream (adoption and diffusion) components of the AI value chain. Second, we summarise the six articles included in the Special Issue ‘AI and labor markets: opening the black box’, distinguishing between contributions focusing on AI exposure, occupations and skill demand; the relationship between AI and automation technologies and their impact on income distribution; and, finally, the effect on organisational structures, management practices, and power dynamics within workplaces. Our analysis emphasises that AI’s employment effects are neither predetermined nor uniform, but shaped by implementation contexts, organisational choices, and institutional frameworks. We find that heterogeneity matters at multiple levels—across countries, sectors, firms, and demographic groups—challenging deterministic narratives and highlighting the need for adaptive policy responses that recognise these asymmetries." (Author's abstract, IAB-Doku) ((en))

Zitatform

Guarascio, Dario, Alessandro Piccirillo & Jelena Reljic (2025): Robots vs. Workers: Evidence From a Meta‐Analysis. In: Journal of Economic Surveys, Jg. 39, H. 5, S. 2254-2271. DOI:10.1111/joes.12699

Abstract

"This study conducts a meta-analysis to assess the effects of robotization on employment and wages, synthesizing the evidence from 33 studies (644 estimates) on employment and a subset of 19 studies (195 estimates) on wages. The results challenge the alarmist narrative about the risk of widespread technological unemployment, suggesting that the overall relationship between robotization and employment or wages is minimal. However, the effects are far from uniform, with adverse outcomes observed in specific contexts, such as the United States, manufacturing sectors, and middle-skilled occupations. The analysis also identifies a publication bias favoring negative wage effects, though correcting for this bias confirms the negligible impact of robotization." (Author's abstract, IAB-Doku) ((en))

Zitatform

Guarascio, Dario & Jelena Reljic (2025): AI and employment in Europe. In: Economics Letters, Jg. 247. DOI:10.1016/j.econlet.2025.112183

Abstract

"This paper contributes to the growing research on AI's labor market impact by presenting novel evidence on the heterogeneous employment effects of AI across EU countries from 2012 to 2022. While concerns persist about AI's disruptive potential, our findings show that occupations more exposed to AI technologies experience stronger employment growth, all else being equal. However, these effects are not uniform across the EU. Positive employment outcomes are concentrated in Innovation Leaders (Belgium, Denmark, Finland, the Netherlands and Sweden) and Strong Innovators (Austria, Cyprus, France, Germany, Ireland and Luxembourg), emphasizing the context-dependent nature of AI's impact. These findings reflect the uneven distribution of innovation capabilities, with a country's innovation system and ‘absorptive capacity’ playing a crucial role in fully harnessing AI's potential for employment (and economic) growth. Ultimately, this research challenges the notion of AI as universally beneficial or harmful, highlighting its asymmetric effects across countries and occupations." (Author's abstract, IAB-Doku, © 2025 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.) ((en))

Zitatform

Guarascio, Dario, Jelena Reljic & Roman Stöllinger (2025): Diverging paths: AI exposure and employment across European regions. In: Structural Change and Economic Dynamics, Jg. 73, S. 11-24. DOI:10.1016/j.strueco.2024.12.010

Abstract

"This study explores exposure to artificial intelligence (AI) technologies and employment patterns in Europe. First, we provide a thorough mapping of European regions focusing on the structural factors—such as sectoral specialisation, R&D capacity, productivity and workforce skills—that may shape diffusion as well as economic and employment effects of AI. To capture these differences, we conduct a cluster analysis which group EU regions in four distinct clusters: high-tech service and capital centres, advanced manufacturing core, southern and eastern periphery. We then discuss potential employment implications of AI in these regions, arguing that while regions with strong innovation systems may experience employment gains as AI complements existing capabilities and production systems, others are likely to face structural barriers that could eventually exacerbate regional disparities in the EU, with peripheral areas losing further ground." (Author's abstract, IAB-Doku, © 2024 The Author(s). Published by Elsevier B.V.) ((en))

Zitatform

Hacker, Winfried (2025): Auswirkungen von KI auf die Nutzer: Erhalten und Fördern der menschlichen Intelligenz bei zunehmendem Einsatz künstlicher Intelligenz - Wozu? Wie? (baua: Fokus), Dortmund, 6 S. DOI:10.21934/baua:fokus20251218

Abstract

"Die Entwicklung der KI verändert die Anforderungen an die menschliche Intelligenz: Denkleistungen können überflüssig werden. Dadurch kann eine arbeitsbedingte Dequalifizierung der Arbeitenden entstehen, denen jedoch die Kontrolle und Korrektur der KI-Ergebnisse obliegt, wofür diese Denkleistungen benötigt werden. Auswege sind die "Zusammenarbeit" von KI und Mensch sowie insbesondere einfache Maßnahmen zum Erhalten der Denkfähigkeit im Arbeitsprozess, die dargestellt werden." (Autorenreferat, IAB-Doku)

Zitatform

Hall, Anja & Ana Santiago Vela (2025): Arbeiten mit Künstlicher Intelligenz, aber auch mit Köpfchen. Anforderungen an Future Skills in der Erwerbsarbeit. In: Berufsbildung in Wissenschaft und Praxis H. 4, S. 21-25.

Abstract

"Künstliche Intelligenz (KI) verändert nicht nur, was wir arbeiten, sondern auch wie. Auf Basis der BIBB/BAuA-Erwerbstätigenbefragung 2024 zeigt der Beitrag die aktuelle Verbreitung von KI auf dem Arbeitsmarkt. KI wird vor allem in kognitiv-analytischen und interaktiven Nichtroutinetätigkeiten genutzt und geht mit Anforderungen an Future Skills wie Probleme lösen, Wissenslücken schließen, kreativ sein oder überzeugen einher. Damit rücken im Kontext von KI neben fachlichen Anforderungen auch überfachliche Kompetenzen stärker in den Fokus. Berufliche Handlungskompetenz ist daher weiterhin gezielt zu fördern." (Autorenreferat, IAB-Doku)

Zitatform

Hammermann, Andrea & Louisa Kürten (2025): Generative KI: Schritt halten durch gezielte Kompetenzentwicklung. (IW-Kurzberichte / Institut der Deutschen Wirtschaft Köln 2025,24), Köln, 3 S.

Abstract

"Der Einsatz von generativer Künstlicher Intelligenz (KI) transformiert die Arbeitswelt in einem rasanten Tempo. Eine wichtige Säule zur Ausschöpfung der möglichen KI-Potenziale sind das Wissen und die Anwendungskompetenz von Beschäftigten. Weiterbildung und das Lernen am Arbeitsplatz gewinnen vor diesem Hintergrund an Bedeutung." (Autorenreferat, IAB-Doku)

Zitatform

Hampole, Menaka, Dimitris Papanikolaou, Lawrence D. W. Schmidt & Bryan Seegmiller (2025): Artificial Intelligence and the Labor Market. (NBER working paper / National Bureau of Economic Research 33509), Cambridge, Mass, 58 S.

Abstract

"We leverage recent advances in NLP to construct measures of workers' task exposure to AI and machine learning technologies over the 2010 to 2023 period that vary across firms and time. Using a theoretical framework that allows for a labor-saving technology to affect worker productivity both directly and indirectly, we show that the impact on wage earnings and employment can be summarized by two statistics. First, labor demand decreases in the average exposure of workers' tasks to AI technologies; second, holding the average exposure constant, labor demand increases in the dispersion of task exposures to AI, as workers shift effort to tasks that are not displaced by AI. Exploiting exogenous variation in our measures based on pre-existing hiring practices across firms, we find empirical support for these predictions, together with a lower demand for skills affected by AI. Overall, we find muted effects of AI on employment due to offsetting effects: highly-exposed occupations experience relatively lower demand compared to less exposed occupations, but the resulting increase in firm productivity increases overall employment across all occupations." (Author's abstract, IAB-Doku) ((en))

Zitatform

Hausman, Naomi, Oren Rigbi & Sarit Weisburd (2025): Generative AI's Impact on Student Achievement and Implications for Worker Productivity. (CESifo working paper 11843), München, 39 S.

Abstract

"Student use of Artificial Intelligence (AI) in higher education is reshaping learning and redefining the skills of future workers. Using student-course data from a top Israeli university, we examine the impact of generative AI tools on academic performance. Comparisons across more and less AI-compatible courses before and after ChatGPT's introduction show that AI availability raises grades, especially for lower-performing students, and compresses the grade distribution, eroding the signal value of grades for employers. Evidence suggests gains in AI-specific human capital but possible losses in traditional human capital, highlighting benefits and costs AI may impose on future workforce productivity." (Author's abstract, IAB-Doku) ((en))

Zitatform

Herzberg-Druker, Efrat (2025): Occupational gender segregation: what can we learn from computer use trends? In: Social forces, S. 1-22. DOI:10.1093/sf/soaf180

Abstract

"This study posits to an intricate interrelation between changes in occupational gender segregation (OGS) and the rise in computer use in the workplace in the United States. I posit that two contrasting mechanisms underpin this relation. Firstly, computerization has contributed to a more balanced gender distribution in certain professions, previously dominated by men, due to a decrease in physical tasks in occupations, thereby reducing OGS. Conversely, in other occupations, heightened computer use has increased Science, Technology, Engineering, and Mathematics (STEM) knowledge requirements, thus restricting women’s integration and reproducing OGS.My empirical analysis, utilizing fixed-effects regression models, lagged models, ordinary least squares (OLS) models, and mediation analysis on a comprehensive dataset of the United States Census, American Community Survey, and Occupational Information Network data, confirms a significant association between computer use and OGS. The physical attributes of occupations and their required STEM knowledge components emerge as critical factors. These contradictory mechanisms—one involving reduced physical demands and the other increased required STEM knowledge—ultimately maintain a stable OGS level." (Author's abstract, IAB-Doku) ((en))

Zitatform

Heyer, Philipp, Kathrin Weis, Sabine Mohr & Wiebke Schmitz (2025): Alter(n) im Betrieb: Stereotype Altersbilder, Fachkräftemangel und digitale Transformation. (BIBB-Report 2025,05), Leverkusen: Verlag Barbara Budrich, 16 S.

Abstract

"Against the backdrop of demographic change, age-appropriate human resources policies are becoming increasingly important. Nevertheless, negative age stereotypes continue to prevail in many firms, hindering the recruitment and further training of older employees – and thus leaving existing skilled labor potential untapped. Based on current data from the establishment survey “BIBB Establishment Panel on Training andCompetence Development,” this BIBB Report analyzes stereotypical images of age in firms as well as company characteristics that promote the employment of older people. Particular attention is given to the role of digital technologies. The results show that the perceptions of older employees vary depending on the industry, firm size, and use of technology. A positive perception is associated with higher employment rates of older people. However, older employees are less strongly represented in firms with above-average use of digital technologies. Based on these findings, it is recommended to counteract age stereotypes, provide targeted further training for older employees, and actively involve them in digital work processes. An age-appropriate human resources policy not only strengthens the supply of skilled workers, but also diversity and, ultimately, the productivity of firms." (Author's abstract, IAB-Doku) ((en))

Zitatform

Humlum, Anders & Emilie Vestergaard (2025): Large Language Models, Small Labor Market Effects. (BFI Working Papers / University of Chicago, Becker Friedman Institute for Research in Economics 2025,56), Chicago, 64 S. DOI:10.2139/ssrn.5219933

Abstract

"We examine the labor market effects of AI chatbots using two large-scale adoption surveys (late 2023 and 2024) covering 11 exposed occupations (25,000 workers, 7,000 workplaces), linked to matched employer-employee data in Denmark. AI chatbots are now widespread —most employers encourage their use, many deploy in-house models, andtraining initiatives are common. These firm-led investments boost adoption, narrow demographic gaps in take-up, enhance workplace utility, and create new job tasks. Yet, despite substantial investments, economic impacts remain minimal. Using difference-in-differences and employer policies as quasi-experimental variation, we estimate precise zeros: AI chatbots have had no significant impact on earnings or recorded hours in any occupation, with confidence intervals ruling out effects larger than 1%. Modest productivity gains (average time savings of 3%), combined with weak wage pass-through, help explain these limited labor market effects. Our findings challenge narratives of imminent labor market transformation due to Generative AI." (Author's abstract, IAB-Doku) ((en))

Zitatform

Högemann, Malte, Laura Hein, Jan-Oliver Britsche & Oliver Thomas (2025): Technostress and generative AI in the workplace: a qualitative analysis of young professionals. In: Frontiers in artificial intelligence, Jg. 8. DOI:10.3389/frai.2025.1728881

Abstract

"Generative artificial intelligence (GenAI) is rapidly diffusing into the workplace and is expected to substantially reshape roles, workflows, and skill requirements, particularly for young professionals as early adopters who are highly exposed to these tools. While GenAI is widely regarded as a means to increase productivity, its adoption may simultaneously introduce new challenges, including various forms of technostress. Drawing on 15 semi-structured interviews with young professionals in research and development (R&D), IT, finance, and marketing in organizations piloting or using GenAI, we conducted a structured qualitative content analysis guided by established technostress dimensions. Our findings indicate that classic technostress dimensions remain relevant but manifest differently across sectors and contexts. Moreover, additional GenAI-specific stressors emerged, such as regulatory and compliance ambiguity, data protection and copyright concerns, perceived dependency, potential skill degradation, doubts about the reliability and controllability of AI outputs, and a shift towards more monitoring and conceptual work. At the same time, participants reported techno-eustress in the form of efficiency gains, learning opportunities, and enhanced intrinsic motivation. Overall, the study extends existing technostress frameworks and underscores the importance of AI literacy, clear organizational governance, and supportive work design to mitigate negative technostress while enabling the productive use of GenAI." (Author's abstract, IAB-Doku) ((en))

Zitatform

Ide, Enrique & Eduard Talamas (2025): The Impact of AI on Global Knowledge Work. (CEPR discussion paper / Centre for Economic Policy Research 20801), London, 34 S.

Abstract

"Artificial Intelligence (AI) is reshaping offshoring and globalization by automating knowledge work and altering trade patterns. We analyze this transformation in a two-region world where firms structure work hierarchically to use knowledge efficiently: the most knowledgeable individuals specialize in problem-solving, while others perform routine work. Before AI, the Advanced Economy specializes in problem-solving services, while the Emerging Economy focuses on routine knowledge work. We model AI as a technology that converts compute into autonomous “AI agents,” which serve as perfect substitutes for humans with a given level of knowledge. Reflecting the concentration of AI infrastructure in advanced economies, we assume that all compute is located in the Advanced Economy. We show that basic AI reduces the Advanced Economy’s net exports of problem-solving services, potentially reversing pre-AI trade patterns. In contrast, sophisticated AI increases the Advanced Economy’s net exports of problem-solving services, reinforcing existing trade patterns. We also examine the effects of restricting AI autonomy, finding that a global restriction redistributes AI’s benefits toward lower-skilled workers, while a regional restriction - such as banning autonomous AI in the Emerging Economy - does little to benefit lower-skilled workers and harms the most knowledgeable individuals in that region. Our results underscore the need for a coordinated global approach to AI regulation." (Author's abstract, IAB-Doku) ((en))

Zitatform

Jaccoud, Florencia (2025): Robots & AI exposure and wage inequality: a within occupation approach. In: Eurasian business review, Jg. 15, H. 4, S. 1035-1090. DOI:10.1007/s40821-025-00306-w

Abstract

"This paper examines the linkages between occupational exposure to recent automation technologies and inequality across 19 European countries. Using data from the European Union Structure of Earnings Survey (EU-SES), a fixed-effects model is employed to assess the association between occupational exposure to artificial intelligence (AI) and to industrial robots–two distinct forms of automation–and within-occupation wage inequality. The analysis reveals that occupations with higher exposure to robots tend to have lower wage inequality, particularly among workers in the lower half of the wage distribution. In contrast, occupations more exposed to AI exhibit greater wage dispersion, especially at the top of the wage distribution. We argue that this disparity arises from differences in how each technology complements individual worker abilities: robot-related tasks often complement routine physical activities, while AI-related tasks tend to amplify the productivity of high-skilled, cognitively intensive work." (Author's abstract, IAB-Doku) ((en))

Zitatform

Jacobs, Arthur & Freddy Heylen (2025): Demographic change, secular stagnation, and inequality: automation as a blessing? In: Journal of demographic economics, Jg. 91, H. 4, S. 508-548. DOI:10.1017/dem.2024.10

Abstract

"We study whether the increased adoption of available automation technologies allows economies to avoid the negative effect of aging on per capita output. We develop a quantitative theory in which firms choose to which extent they automate in response to a declining workforce and rising old-age dependency. An important element in our model is the integration of two capital types: automation capital that acts as a substitute to human labor, and traditional capital that is a complement to labor. Empirically, our model's predictions largely match data regarding automation (robotization) density across OECD countries. Simulating the model, we find that aging-induced automation only partially compensates the negative growth effect of aging in the absence of technical progress in automation technology. One reason is that automated tasks are no perfect substitutes for non-automated tasks. A second reason is that automation raises the interest rate and thus inhibits positive behavioral reactions to aging (later retirement and investment in human capital). Moreover, increased automation generates a falling net labor share of income and rising welfare inequality. We evaluate alternative policy responses to cope with this inequality." (Author's abstract, IAB-Doku) ((en))

Zitatform

Janssen, Simon, Christina Langer, Markus Nagler, Nicholas Rounding, Michael Stops & Simon Wiederhold (2025): Wie lässt sich die Nachfrage nach KI- und anderen Kompetenzen auf dem Arbeitsmarkt besser messen? (ROA external reports / Researchcentrum voor Onderwijs en Arbeidsmarkt (Maastricht) 10 ai:conomics policybrief), Maastricht, 6 S.

Abstract

"Eine umfangreiche Forschungsliteratur zeigt, dass der technologische Wandel erhebliche Auswirkungen auf die Arbeitsmärkte hat, da moderne digitale Technologien die Nachfrage nach bestimmten Kompetenzen verändern. Zum einen können neue Technologien einige menschliche Tätigkeiten ersetzen. Zum anderen Seite können sie neue Tätigkeiten schaffen oder ergänzen (Acemoglu et al., 2015; Acemoglu & Restrepo, 2018, 2019, 2020). Mit der starken Verbreitung Künstlicher Intelligenz in den letzten Jahren gewinnen bestimmte Fragen in der öffentlichen Diskussion und der Forschung zunehmend an Bedeutung: Wächst die Arbeitsnachfrage nach KI-Kompetenzen auch auf dem deutschen Arbeitsmarkt? Führt die steigende Nachfrage nach KI-Kompetenzen dazu, dass andere Kompetenzen – bei niedrig-, mittel- und hochqualifizierten Arbeitskräften – weniger gefragt sind? Ziel dieses Forschungsprojekts ist es, eine belastbare Datengrundlage zu schaffen, um solche Fragen in Zukunft fundierter beantworten zu können. Die Entwicklungen bei generativer Künstlicher Intelligenz, insbesondere von Tools wie ChatGPT, hat die Diskussion über die Auswirkungen von KI auf den Arbeitsmarkt sowohl in der Wissenschaft als auch in der öffentlichen Debatte und in der Politik deutlich verstärkt. Während Computer und Software die Arbeitswelt durch die präzisere und effizientere Ausführung routinemäßiger Aufgaben verändert haben, können moderne KI-Systeme nun komplexe, nichtroutinemäßige Aufgaben übernehmen, ohne auf detaillierte Anweisungen oder wiederholende Regeln angewiesen zu sein (Brynjolfsson et al., 2025). Infolgedessen sehen viele das produktive Potenzial dieser neuen Technologie optimistisch. Andere hingegen befürchten, dass KI die Arbeitsmärkte disruptiv verändern könnte." (Autorenreferat, IAB-Doku)

Zitatform

Janssen, Simon, Christina Langer, Markus Nagler, Nicholas Rounding, Michael Stops & Simon Wiederhold (2025): How can we better measure the demand for AI and other skills on the labour market? (ROA external reports / Researchcentrum voor Onderwijs en Arbeidsmarkt (Maastricht) 10 ai:conomics policybrief), Maastricht, 5 S.

Abstract

"A large body of research literature shows that technological change has a significant impact on labour markets, as modern digital technologies are changing the demand for certain skills. On the one hand, new technologies can replace some human activities. On the other hand, they can create or complement new activities (Acemoglu et al., 2015; Acemoglu & Restrepo, 2018, 2019, 2020). With the proliferation of artificial intelligence (AI) in recent years, certain questions are becoming increasingly important in public debate and research: 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 aim of this research project is to create a reliable data basis in order to be able to answer such questions in a more informed way in the future. Developments in generative AI, particularly tools such as ChatGPT, have significantly intensified the discussion about the impact of AI on the labour market, both in academia and in public debate and policy. While computers and software have transformed the world of work by performing routine tasks more precisely and efficiently, modern AI systems can now take on complex, non-routine tasks without relying on detailed instructions or repetitive rules (Brynjolfsson et al., 2025). As a result, many are optimistic about the productive potential of this new technology. Others, however, fear that AI could disrupt labour markets. In the course of the intensive scientific and public debate on AI, there is a growing body of literature that deals with the effects of AI on labour markets. These initially focus on specific occupations such as call centre workers (Brynjolfsson et al., 2025, Dijksman et al., 2024), consultants (Dell’ et al., 2023), writers or developers (Peng et al., 2023). However, a major challenge is to measure how the demand for and supply of skills has changed in the wake of the emergence of AI." (Autorenreferat, IAB-Doku)

Zitatform

Janssen, Simon, Michael Stops, Sanne Steens, Pelin Özgül, Nicholas Rounding, Sander Dijksman, Raymond Montizaan, Mark Levels, Didier Fourage, Danique Eijkenboom, Evie Graus & Marie-Christine Fregin (2025): Artificial intelligence in the workplace: insights into the transformation of customer services. In: IAB-Forum H. 22.04.2025, 2025-04-22. DOI:10.48720/IAB.FOO.20250422.01

Abstract

"How does the use of artificial intelligence in training affect employee productivity? These and other questions were investigated as part of the long-term research project “ai:conomics” using company data from various large European companies. Initial results suggest that AI can have a positive impact on employee productivity, especially for new employees." (Author's abstract, IAB-Doku) ((en))

Zitatform

Khabirpour, Neysan, Lorenzo Pelizzari, Jan Limbers, Markus Hoch, Philipp Kreuzer, Matteo Richiardi, Patryk Bronka & Costanza Pagnini (2025): Overlapping crises (re)shaping the future of regional labour markets [OVERLAP]. Main Report. Luxembourg: ESPON 2030, 101 S.

Abstract

"Europe’s labour market is entering a decade in which structural forces converge and pull unevenly on every region. First, the demographic change is steadily thinning the labour supply: by 2050, the EU’s labour force is set to decrease by 35 million people. (Secondly) This demographic transition comes at a time when Member States are increasing their efforts to achieve the decarbonization targets, and (thirdly) when they are ramping up investments to digitalise the economy. While the digital transition is accelerating demand for specialised skills faster than workers can acquire them, the green transition suggests both disruption and expansion. To deliver the REPowerEU targets, the Commission estimates that more than 3.5 million additional jobs will be needed by 2030. Explained very shortly, these interacting factors may amplify long-standing territorial disparities in age structure, industrial fabric and human-capital endowment. Understanding where labour will transform and where new demand will arise is therefore indispensable. It is precisely this spatial intelligence that the OVERLAP project supplies—by charting the possible employment trajectories of every NUTS-3 labour market – for the 2035 perspective - under a varied set of assumptions, driven by policy or shock. This is done within a scenario-driven exercise, covering ageing, green ambition and digital diffusion. In doing so, as a forward-looking exercise, the study equips policymakers with the granular evidence needed to anticipate potential shortages, target up- and reskilling investments, and steer and match transition funding to address local needs and the regions that need it most. The study starts from two overarching objectives: Compile a granular portrait of Europe’s regional labour markets by tracing demographic dynamics and their (possible) implications for employment trends, at NUTS-3 level, out to 2035. Gauge how major drivers—including population change and the twin digital-green transition—may reshape labour demand under a range of forward-looking (possible) scenarios, out to 2035. From these aims, flow the main guiding research questions: which territories and sectors are set to gain or lose employment as ageing, automation and decarbonisation unfold simultaneously? And what policy mixes can cushion vulnerable regions while helping them capture new growth niches? Addressing these questions across the ESPON space—i.e. all EU Member States plus Iceland, Liechtenstein, Norway and Switzerland—requires a geography-sensitive lens; hence results are mapped down to individual NUTS-3 regions. To deliver evidence at that resolution, the project combined a dual analytical architecture. Quantify and regionalise macro-trends: the top-down stream employs the DINOS dynamic input-output model (developed by PROGNOS) to translate demographic, technological and climate-policy shocks into sectoral employment and wage shifts, then regionalises these outputs to the full NUTS-3 grid. The macro-level modelling strategy begins with national economic aggregates, traces broad structural trends across industries, and subsequently disaggregates the resulting labour-market effects to individual regions. By working from the “whole economy” downward, this framework captures systemic interactions—such as supply-chain spill-overs—beyond the reach of purely regional models. Provide a micro-analytical perspective: in parallel, the bottom-up stream extends the SimPaths dynamic microsimulation platform—already validated for the United Kingdom as a baseline —to Greece, Hungary, Italy and, enriching the macro picture with individual life-course trajectories. This novel, regional, micro-analytical framework sheds additional light on the distributional impact of the ongoing economic and social transformations, going beyond the broad picture and simplified assumptions that had to be made in the top-down approach (macro-analyses). The dynamic framework integrates a defining feature in every simulated period: inputs from a static tax-benefit calculator (EURO-MOD), hence allowing to study to what extent tax and benefit systems can smooth out transitional dynamics. However, it is important to highlight from the onset, that this study should not be perceived as a crystal globe, as it does not cover all possible shocks or situations, but acts upon the accumulated knowledge, in order to provide some modelled scenarios that are aimed at informing and opening the forward-looking strategies, with a pre-emptive component." (Author's abstract, IAB-Doku) ((en))

Zitatform

Klonek, Florian & Sharon Parker (2025): Does AI at Work Increase Stress? Text Mining Social Media About Human–AI Team Processes and AI Control. In: Journal of organizational behavior, S. 1-15. DOI:10.1002/job.70000

Abstract

"With rising use of artificial intelligence (AI) in organizations, alongside increasing mental health issues, we seek to understand how AI use affects human stress. Drawing on the automation–augmentation perspective, we propose that AI control over decision-making thwarts human autonomy and thus contributes to stress. Drawing on models of teamwork and augmentation, we expect that human–AI team processes (i.e., transition, action, and interpersonal processes) help people meet their goals and reduce stress. Finally, we argue that human–AI team processes provide an important social resource, which buffers the stress-enhancing role of AI control. To test our hypotheses, we analyzed over 2700 tweets. Using a trained large language model, validated against human ratings, we indexed key measures. Results confirm that high AI control was associated with increased stress, whereas human–AI team processes were associated with decreased stress. In support of the moderation hypothesis, two human–AI team processes (action and interpersonal) helped further reduce the stress-enhancing effect of AI control. We discuss implications for work design theory and the importance of regulating levels of AI control to protect workers' mental health." (Author's abstract, IAB-Doku) ((en))

Zitatform

Koch, Christian & Michael Stops (2025): KI Navigator #10: Wie KI dem Arbeitsmarkt hilft. In: Heise online, 2025-03-14.

Abstract

"Stellenanzeigen können viel über den Wandel des Arbeitsmarkts verraten. Künstliche Intelligenz hilft dabei, diese Daten zu interpretieren."

Zitatform

Kowalik, Zuzanna, Piotr Lewandowski, Tomasz Geodecki & Maciej Grodzicki (2025): Automation in shared service centres: Implications for skills and autonomy. In: The Economic and Labour Relations Review, Jg. 36, H. 2, S. 563-581. DOI:10.1017/elr.2025.10026

Abstract

"The offshoring-fueled growth of the Central and Eastern European business services sector gave rise to shared service centers (SSCs) – quasi-autonomous entities providing routine-intensive tasks for the central organization. The advent of technologies such as intelligent process automation, robotic process automation, and artificial intelligence jeopardises SSCs’ employment model, necessitating workers’ skills adaptation. The study challenges the deskilling hypothesis and reveals that automation in the Polish SSCs is conducive to upskilling and worker autonomy. Drawing on 31 in-depth interviews, we highlight the negotiated nature of automation processes shaped by interactions between headquarters, SSCs, and their workers. Workers actively participated in automation processes, eliminating the most mundane tasks. This resulted in upskilling, higher job satisfaction, and empowerment. Yet, this phenomenon heavily depends upon the fact that automation is triggered by labor shortages, which limit the expansion of SSCs. This situation encourages companies to leverage the specific expertise entrenched in their existing workforce. The study underscores the importance of fostering employee-driven automation and upskilling initiatives for overall job satisfaction and quality." (Author's abstract, IAB-Doku) ((en))

Zitatform

Krzywdzinski, Martin, Daniel Schneiß & Andrea Sperling (2025): Between control and participation: The politics of algorithmic management. In: New Technology, Work and Employment, Jg. 40, H. 1, S. 60-80. DOI:10.1111/ntwe.12293

Abstract

"Understanding the role of human management is crucial for the debate over algorithmic management—to date limited to studies on the platform economy. This qualitative case study in logistics reconstructs the actor constellations (managers, engineers, data scientists and workers) and negotiation processes in different phases of algorithmic management. It offers two major contributions to the literature: (1) a process model distinguishing three phases: goal formation, data production and data analysis, which is used to analyse (2) the politics of algorithmic management in conventional workplaces, which differ significantly from platform companies. The article goes beyond surveillance to elucidate the role of the regulatory framework, various actors' knowledge contributions to the algorithmic management system, and the power relations resulting therefrom. While the managerial goals in the examined case were not oriented towards a surveillance regime, the outcome was nevertheless a centralisation of knowledge and disempowerment of workers." (Author's abstract, IAB-Doku) ((en))

Zitatform

Kuhlmann, Martin, Britta Matthes & Stefan Theuer (2025): Das Produktionsmodell der deutschen Automobilindustrie auf dem Prüfstand. Arbeitsstrukturen und Arbeitsanforderungen in Montagewerken im Wandel? (SOFI-Impulspapier), Göttingen, 6 S.

Abstract

"Das in den 1980er-Jahren etablierte Produktionsmodell der deutschen Automobilhersteller lässt sich beschreiben als innovations- und exportorientierte Produktion qualitativ hochwertiger Produkte auf Basis qualifizierter Arbeit, guter Bezahlung und hoher Beschäftigungssicherheit sowie starken gewerkschaftlichen Interessenvertretungen. Politische Vorgaben, wie die Umstellung auf die Produktion von Elektroautos, veränderte Wettbewerbsbedingungen sowie die weiter voranschreitende Digitalisierung haben dazu geführt, dass dieses Produktionsmodell derzeit auf dem Prüfstand steht. Getrieben durch aufkommende Zweifel an der technologischen Überlegenheit deutscher Automobilhersteller und Nachfrageschwächen beim Übergang auf Elektromobilität ist die Unsicherheit in der Branche gegenwärtig groß. In einem laufenden Forschungsprojekt untersuchen wir, inwiefern sich durch die Produktion von Elektroautos und die fortschreitende Digitalisierung Arbeitsstrukturen und Arbeitsanforderungen in den Endmontagewerken der deutschen Automobilhersteller verändert haben und ob sich arbeitsbezogen ein Wandel des deutschen Produktionsmodells abzeichnet." (Autorenreferat, IAB-Doku)

Zitatform

Kuhn, Sarah & Holger Seibert (2025): Digitale Ersetzbarkeit im stationären Einzelhandel und im Onlinehandel. Gastbeitrag. In: ArbeitGestalten Beratungsgesellschaft mbH (Hrsg.) (2025): Kassensturz. Daten, Fakten und Erfahrungen aus der Arbeitswelt des Berliner Einzelhandels, S. 29-31, 2025-09-30.

Abstract

"Diese Analyse zeigt, dass nicht alle Beschäftigten im stationären Einzelhandel und im Onlinehandel gleichermaßen von der digitalen Transformation betroffen sind bzw. sein werden. Wichtig zu betonen ist, dass es hier um Potenziale technischer Ersetzbarkeit geht. Ob und inwiefern die technischen Möglichkeiten auch tatsächlich umgesetzt werden, hängt von verschiedenen Faktoren ab. Es kann Gründe geben, die gegen eine tatsächliche Substituierung sprechen, beispielsweise weil eine Umstellung zu komplex wäre oder ethische Bedenken dem entgegenstehen. Unstrittig ist jedoch, dass auf der einen Seite einige Tätigkeiten durch die Digitalisierung wegfallen bzw. automatisiert werden, andererseits aber auch neue Tätigkeiten und Berufe entstehen. Daher kann ein hohes Substituierungspotenzial als Indikator für einen Wandel der Arbeitswelt gesehen werden." (Autorenreferat, IAB-Doku)

Zitatform

Kuhn, Sarah & Holger Seibert (2025): Digitalisierung der Arbeitswelt: Durch künstliche Intelligenz sind inzwischen auch viele Expertentätigkeiten ersetzbar. (IAB-Regional. Berichte und Analysen aus dem Regionalen Forschungsnetz. IAB Berlin-Brandenburg 01/2025), 34 S. DOI:10.48720/IAB.REBB.2501

Abstract

"Durch neue digitale Technologien verändert sich der deutsche Arbeitsmarkt. Dies gilt besonders für das Ausmaß, in dem Berufe aktuell potenziell durch den Einsatz von Computern oder computergesteuerten Maschinen ersetzbar sind, dem so genannten Substituierbarkeitspotenzial. Es beschreibt, welcher Anteil an Tätigkeiten in einem Beruf schon heute durch den Einsatz moderner Technologien ersetzt werden könnte. Nach wie vor ist zwar das Substituierbarkeitspotenzial bei den Helfer*innen- und Fachkraftberufen am höchsten. Am stärksten gestiegen ist das Potenzial jedoch bei den Expert*innenberufen (u. a. durch generative Künstliche Intelligenz). Besonders bei den IT- und naturwissenschaftlichen Dienstleistungsberufen sind hohe Zuwachsraten zwischen 2019 und 2022 zu verzeichnen. Der vorliegende Beitrag fokussiert sich auf den Arbeitsmarkt in Brandenburg und Berlin. Wichtig zu betonen ist, dass es hier um Potenziale technischer Ersetzbarkeit geht. Ob und inwiefern die technischen Möglichkeiten auch tatsächlich umgesetzt werden, steht nicht fest. Es kann Gründe geben, die gegen eine tatsächliche Substituierung sprechen, beispielsweise weil eine Umstellung zu komplex wäre oder ethische Bedenken dem entgegenstehen. Unstrittig ist jedoch, dass auf der einen Seite einige Tätigkeiten durch die Digitalisierung wegfallen bzw. automatisiert werden, andererseits aber auch neue Tätigkeiten und Berufe entstehen. Daher kann ein hohes Substituierungspotenzial als Indikator für einen Wandel der Arbeitswelt gesehen werden." (Autorenreferat, IAB-Doku)

Zitatform

Kunze, Florian, Carolina Opitz & Ann Sophie Lauterbach (2025): Konstanzer KI-Studie 2025: Die Nutzung von Künstlicher Intelligenz in der Arbeitswelt steigt, Ungleichheiten in der Wahrnehmung bleiben weiterhin bestehen. Ergebnisbericht Juli 2025. Konstanz: KOPS Universität Konstanz, 8 S.

Abstract

"Die Nutzung von KI in der Arbeitswelt hat innerhalb eines Jahres deutlich zugenommen – gleichzeitig bleiben erhebliche Unterschiede zwischen Berufsgruppen, Bildungsniveaus und Unternehmen bestehen. In der zweiten Welle der Konstanzer KI-Studie berichten 35?% der Befragten von KI-Nutzung im Arbeitsalltag, ein Anstieg um 11 Prozentpunkte gegenüber dem Vorjahr. Trotz dieses Wachstums bleibt die Unsicherheit hoch: Ein Drittel der Beschäftigten kann weiterhin nicht einschätzen, welche Folgen KI für die eigene Arbeit haben wird. Zugleich wird der gesellschaftliche Einfluss von Automatisierung deutlich bedrohlicher wahrgenommen als die persönliche Betroffenheit. Besonders stark ist der Nutzungszuwachs in wissensintensiven Berufen, während produktionsnahe Tätigkeiten kaum aufholen. Auch die Kluft zwischen Bildungsgruppen bleibt bestehen: Beschäftigte mit hohem Bildungsabschluss nutzen KI mehr als dreimal so häufig wie jene mit niedrigem Abschluss. Zwar steigt die Bereitschaft zur Weiterbildung in allen Gruppen, strukturelle Hürden scheinen jedoch eine Angleichung zu verhindern. Auf Ebene der Organisationen verlaufen die Entwicklungen deutlich langsamer als auf individueller Ebene. Vor allem große Unternehmen investieren zunehmend in Weiterbildung und Führungskommunikation, während kleinere Organisationen kaum Veränderungen zeigen. Die Ergebnisse zeigen deutlich, dass KI ihr Potenzial nicht gleichmäßig entfaltet, sondern bestehende strukturelle Ungleichheiten eher verstärkt. Nach wie vor besteht die reale Gefahr, dass sich bestimmte Beschäftigtengruppen zunehmend vom technologischen Fortschritt abkoppeln, weil ihnen der Zugang zu KI-Nutzung, Weiterbildungsangeboten und betrieblicher Unterstützung fehlt. Daraus ergibt sich ein klarer Handlungsauftrag an Wirtschaft, Politik und Bildungseinrichtungen, um Teilhabechancen gezielt zu fördern und einer wachsenden sozialen Spaltung frühzeitig entgegenzuwirken." (Textauszug, IAB-Doku)

Abstract

"This report examines the potential for generative AI – tools that generate text, images, video or audio, such as ChatGPT, Copilot and Midjourney – to help SMEs address labour and skill needs. It presents evidence from a representative 2024 OECD survey of over 5 000 SMEs in Austria, Canada, Germany, Ireland, Japan, Korea and the United Kingdom, on how SMEs use generative AI, how its use may be helping to address labour and skill needs, and how SMEs are preparing employees to use generative AI. The survey shows that generative AI is in use in 31% of SMEs. SMEs report that generative AI improves performance, helps compensate for skill gaps and labour shortages, and increases the need for highly-skilled workers. SMEs have concerns about copyright, legal and regulatory issues, though negative attitudes towards generative AI are rare. The findings highlight the promise of generative AI but also the need for structured policy support to close digital and skills gaps between SMEs and larger firms and to ensure that any gains from generative AI are broadly shared across the economy and the workforce." (Author's abstract, IAB-Doku) ((en))

Zitatform

Leitner, Sebastian & Stella Sophie Zilian (2025): Digitalisation of jobs and gender-age segregation in digital tasks: Cross-country evidence based on ESJS2 data. (WIIW working paper 269), Wien, 37 S.

Abstract

"This paper addresses the disproportional effects of digitalisation across age by investigating (i) within-job age segregation in tasks by digital intensity; (ii) within-job age disparities in digital upskilling; (iii) age inequalities in wage returns to digital job tasks; and (iv) the role of gender in this age segregation and inequalities. The analysis is based on data of Cedefop's second wave of the European Skills and Jobs Survey (ESJS2), conducted in 2021. First results of the analysis show that even when controlling for occupation-industry job pairs apart from using other explanatory variables, age segregation and gender gaps are prevalent in the case of digital skill intensity of tasks performed in the jobs of employees, though not in the case of digital upskilling via training measures. Applying the same appropriate controls, we also find that higher within-job digital skill intensity is associated with higher hourly wages. Gender wage gaps are sizable across all skill intensity categories in addition to widening in older age groups." (Author's abstract, IAB-Doku) ((en))

Zitatform

Lewandowski, Piotr & Wojciech Szymczak (2025): Automation, Trade Unions and Atypical Employment. In: Industrial Relations, S. 1-19. DOI:10.1111/irel.70017

Abstract

"We study the effect of automation technologies—industrial robots, software and databases—on the incidence of involuntary atypical employment in 13 EU countries between 2006 and 2018. Robots do not affect the total employment rate but significantly increase the involuntary atypical employment share, mainly through fixed-term work. Software and databases increase total employment and are neutral for atypical employment. Higher trade union density mitigates the robots' impact on atypical employment, while employment protection legislation plays no role. Using historical decompositions, we attribute 1–2 percentage points of a 15% average atypical employment share in our sample to automation." (Author's abstract, IAB-Doku, Published by arrangement with John Wiley & Sons) ((en))

Zitatform

Logchies, Thomas, Tom Coupé & W. Robert Reed (2025): Do robots decrease humans’ wages? In: Applied Economics Letters, S. 1-5. DOI:10.1080/13504851.2025.2466748

Abstract

"While there are studies that show a positive or negative impact of robots on wages, a meta-analysis of 2,586 estimates from 52 studies in this paper finds that when one looks at the literature as a whole, there is no clear evidence of a sizable impact of robots on wages." (Author's abstract, IAB-Doku) ((en))

Zitatform

Lott, Yvonne & Andreas Hövermann (2025): Computer Use and Digital Frustration in German Workplaces: Is There a Gendered Part-Time Gap? In: Work, Employment and Society, Jg. 39, H. 6, S. 1440-1462. DOI:10.1177/09500170251351265

Abstract

"The digital transformation may disproportionately disadvantage female part-time workers, as they are affected by the flexibility stigma and career penalties. In this article, we ask: Is there a gendered part-time gap in work-related computer use and digital frustration in Germany? Latent class analysis and multivariate analysis, based on data from Wave 12 (2019/20) of the German National Educational Panel Study (NEPS) Starting Cohort 6 – Adults, showed that women – and part-time working women in particular – were less likely than men to be classified as ‘advanced users’. Furthermore, part-time working women felt least well prepared for using networked digital technologies at work and were thus more at risk of experiencing digital frustration. These findings suggest that the triadic association between technology, power and masculinity postulated by feminist technology theory should be extended to include full-time work." (Author's abstract, IAB-Doku) ((en))

Zitatform

Lábaj, Martin, Tomáš Oleš & Gabriel Procházka (2025): Impact of robots and artificial intelligence on labor and skill demand: evidence from the UK. In: Eurasian business review, Jg. 15, H. 4, S. 953-1001. DOI:10.1007/s40821-025-00314-w

Abstract

"Over the past four decades, automation technologies have replaced routine tasks performed by medium-skilled workers, and contributed to increased labor market polarization. With the advent of artificial intelligence, this dynamic may have shifted, extending task substitution to non-routine tasks performed by high-skilled workers. Using textual analysis and descriptions of technology found in patent texts, we construct novel occupational exposures to robot and artificial intelligence technologies. These occupational exposures are then used to analyze changes in labor and skill demand over the last decade in the United Kingdom. We find that the middle part of the income distribution is primarily exposed to robot technology, while exposure to artificial intelligence increases monotonically across income percentiles. Second, we find that exposure to robots is strongest among high school dropouts and declines monotonically with education, while artificial intelligence automation has a limited impact on the same workers, with a pronounced exposure among college graduates. Third, our findings suggest asymmetric effects of automation technologies across skill groups. Robot automation reduces demand for low-skilled workers, while AI technology shifts demand away from high-skilled workers, with the direct effects consistently negative despite the presence of several compensating mechanisms. Fourth, despite significant effects on wage bill, we find no robust relationship between automation exposure and changes in the employment-to-population ratio. Finally, a joint estimation of the effects of robot and AI automation shows that robot automation is positively associated with an increase in demand for skilled workers, while AI automation is weakly associated with a decrease in demand for skilled workers." (Author's abstract, IAB-Doku) ((en))

Zitatform

Lühr, Thomas & Tobias Kämpf (2025): Bots im Büro. Künstliche Intelligenz und der Wandel von Angestelltenarbeit in der digitalen Transformation. (Hans-Böckler-Stiftung. Study 494), Düsseldorf: Hans-Böckler-Stiftung, Düsseldorf, 98 S.

Abstract

"Mit der digitalen Transformation kommt es zu einem Schub in der Automatisierung von Arbeit. Die Einführung von Künstlicher Intelligenz führt zur grundlegenden Restrukturierung der Arbeitsinhalte und -prozesse im Büro. Damit gehen nicht nur Risiken von Funktionsverlusten bis hin zum Verlust des Arbeitsplatzes einher, sondern auch neue Machtpotenziale. Diese prägen das Bewusstsein der Angestellten wesentlich. Künstliche Intelligenz funktioniert nicht ohne Mitbestimmung - mit Mitbestimmung ergeben sich neue Ansatzpunkte für eine arbeitspolitische Vorwärtsstrategie. Die vorliegende Studie nimmt eine empirisch gestützte Analyse der Potenziale vor, die der Automatisierungsschub für die Beschäftigten und ihre Interessenvertretungen tatsächlich bietet." (Autorenreferat, IAB-Doku)

Zitatform

Machovec, Christine, Michael Rieley & Emily Rolen (2025): Incorporating AI impacts in BLS employment projections: occupational case studies. In: Monthly labor review H. February. DOI:10.21916/mlr.2025.1

Abstract

"In the last few years, artificial intelligence (AI) has advanced rapidly, finding growing applications across industries and occupations. This development has generated interest in how the U.S. Bureau of Labor Statistics assesses and incorporates AI’s potential labor market impacts in its employment projections. In this article, we explain the Bureau’s approach to this type of projections work, illustrating it with several occupational case studies based on research done for the 2023–33 projections cycle. The case studies focus on selected occupations in the computer, legal, business and financial, and architecture and engineering occupational groups." (Author's abstract, IAB-Doku) ((en))

Zitatform

Mak, Eric & Aloysius Siow (2025): Occupational Choice, Matching, and Earnings Inequality. In: Journal of Political Economy, Jg. 133, H. 1, S. 355-383. DOI:10.1086/732530

Abstract

"We combine classic occupational choice (Roy, 1951) and frictionless matching (Sattinger, 1979) to explain earnings by occupation and firm in a way that is consistent with the double assignment. In our model, within-firm inequality is globally non-zero whenever there is asymmetry in the revenue function or the occupational skill distribution across occupations. Occupational earnings overlap each other, and unlike the Roy Model, the distributions of potential earnings are endogenous. In line with recent empirical findings on earning decomposition, skill-biased technical change (SBTC)increases within-firm inequality mostly among high-wage firms and not among low-wagefirms." (Author's abstract, IAB-Doku) ((en))

Zitatform

Malin, Christine, Jürgen Fleiß, Renate Ortlieb & Stefan Thalmann (2025): Rejected by an AI? Comparing job applicants’ fairness perceptions of artificial intelligence and humans in personnel selection. In: Frontiers in artificial intelligence, Jg. 8. DOI:10.3389/frai.2025.1671997

Abstract

"Introduction: Artificial intelligence (AI) transforms personnel selection, but the application of AI raises fairness concerns and aversion towards AI. Although job applicants may perceive the selection process as fairer when they receive an explanation for the decision, scientific knowledge about AI-related fairness perceptions in this setting is limited. This paper investigates how job applicants perceive fairness of an AI-based personnel selection process considering explanations provided. Methods: The hypotheses are based on a theoretical framework about fairness and literature on algorithm aversion. Data were collected through a vignette-style method focusing on four personnel selection scenarios (n = 921). Results: We show that provided explanations increase job applicants ’ perceptions of outcome fairness, process fairness, interpersonal treatment, and recommendation intention, irrespective of the decision being made by an AI or human. Discussion: We provide conclusions for algorithmic decision-making and discuss factors that need to be considered when adopting and designing AI so that AI is perceived as fair." (Author's abstract, IAB-Doku) ((en))

Zitatform

Marczuk, Anna, Frank Multrus, Thomas Hinz & Susanne Strauss (2025): Künstliche Intelligenz (KI) im Studienalltag: Einschätzungen von Studierenden zum Einsatz von KI an deutschen Hochschulen. (DZHW-Brief 2025,02), Hannover, 15 S. DOI:10.34878/2025.02.dzhw_brief

Abstract

"Die Mehrheit der Studierenden nutzt im Wintersemester 2024/2025 KI im Studium und kennt deren Funktionsweise relativ gut. ChatGPT ist das meistgenutzte KI-Tool, dessen Nutzung seit 2023 deutlich angestiegen ist. Studierende verwenden KI am häufigsten für die Einführung in ein Thema und für Textverarbeitungen, deutlich seltener für Literaturrecherchen oder Datenanalysen. Die Mehrheit der Studierenden gibt an, dass KI die Erledigung von Aufgaben, die keinen Spaß machen oder schwierig sind, beschleunigt oder erleichtert. Seltener sind Studierende der Ansicht, dass KI die Studienleistungen verbessert. Studierende stehen KI auch kritisch gegenüber, insbesondere wegen ihrer Fehleranfälligkeit und des Risikos, von ihr abhängig zu werden. Studierende, die KI häufig nutzen, sind gegenüber KI ähnlich kritisch wie Studierende, die sie seltener nutzen. Der Einsatz von Learning Analytics wird von Studierenden eher befürwortet, wenn sie selbst dadurch unterstützt werden (etwa durch Kurs- und Literaturempfehlungen), weniger zur Unterstützung von Lehrenden (etwa bei der Benotung) oder der Hochschulverwaltung (etwa für die Studienabbruchprävention). Studierende erleben eher selten eine Unterstützung der Hochschulen bei der Nutzung von KI im Studium. An einigen Hochschulen berichten sie von Richtlinien zur Nutzung, seltener sind Schulungsangebote oder eine Integration in die Lehre. Studierende wünschen sich KI-Unterstützung beim Verfassen von Hausarbeiten, während der Einsatz durch Lehrende zur Benotung oder als Ersatz für Lerngruppen (automatisierte Lernbuddys) skeptisch gesehen wird. Eine Teildigitalisierung von Lehrveranstaltungen (Mischung aus Präsenz und online) ist für Studierende attraktiver als reine Präsenz- oder gar reine Onlineveranstaltungen." (Autorenreferat, IAB-Doku)

Zitatform

Matthes, Britta (2025): KI und Beratung. In: Dvb-Forum, Jg. 64, H. 1, S. 17-22., 2025-02-14.

Abstract

"Wie KI und andere digitale Technologien den Arbeitsmarkt verändern: Am IAB werden wir immer wieder danach gefragt, welche Berufe angesichts der rasanten technologischen Entwicklungen der letzten Jahre überhaupt noch Zukunft haben. Bislang hat man zur Beantwortung dieser Frage Prognosen zu Rate gezogen. Hier wurde anfangs – unter Berücksichtigung verschiedener relativ stabiler Faktoren wie dem Erwerbspersonenpotenzial, der wirtschaftlichen Entwicklung oder der zu erwartenden Migration – hochgerechnet, wie sich die Zahl der Berufsanfänger auf die verschiedenen Berufe und Qualifikationsniveaus verteilt, wenn die Entwicklung sich wie in der Vergangenheit fortsetzen würde. Schon früh wurde jedoch deutlich, dass diese Faktoren weniger stabil sind als ursprünglich angenommen. Um diese Dynamik zu berücksichtigen, wurde dieser Ansatz erweitert, indem nunmehr Projektionen erstellt werden. Dazu werden Annahmen über die Folgen bestimmter, äußerst wahrscheinlicher Ereignisse oder Verhaltensweisen getroffen, für die sich (noch) keine langfristige Zahlenbasis finden lässt. So gibt die QuBe-Projektion einen langfristigen Überblick über die voraussichtliche Entwicklung des Arbeitskräftebedarfs und -angebotes nach Qualifikationen und Berufen unter einer Reihe von Annahmen über zum Beispiel die Folgen des Klimawandels oder den Ausbau der ökologischen Landwirtschaft. Außerdem werden anhand von Abweichungen zwischen diesem Basismodell und Szenarien die absehbaren Folgen bestimmter Vorhaben oder Ereignisse, wie zum Beispiel der Maßnahmen zur Energie- und -Mobilitätswende abgeschätzt (https://www.bibb.de/de/202333.php). Allerdings sind diese Modelle sehr komplex und es stellt sich die Frage, inwieweit solche Projektionen für die Bildungs- und Berufsberatung einzelner Personen sinnvoll genutzt werden können. Hinzu kommt derzeit, dass die technologische Entwicklung derart schnell voranschreitet, dass verstärkt mit Umwälzungen auf dem Arbeitsmarkt gerechnet werden muss, die auch altbekannte Zusammenhänge in Frage stellen könnten. Für die einzelne Person steht die Frage im Raum, mit welchen Konsequenzen sie selbst rechnen muss, wenn neue Technologien zum Einsatz kommen: Reicht es aus, sich auf den aktuellen Wissensstand im eigenen Beruf zu bringen? Womit sollte man sich konkret beschäftigen, um den Anforderungen des Berufes weiterhin gewachsen zu sein? Ist es zielführender, sich beruflich neu zu orientieren?" (Textauszug, IAB-Doku, © wbv)

Zitatform

Matthes, Britta (2025): Wie KI Berufe verändert und Chancen für Menschen mit Behinderungen schafft. In: Die Berufliche Rehabilitation, Jg. 39, H. 1, S. 6-15., 2025-04-04.

Abstract

"Es ist absehbar, dass die rasanten technologischen Entwicklungen der letzten Jahre, insbesondere die enorme Steigerung der Rechenleistung und die Entwicklung selbstlernender algorithmischer Systeme, die heute allgemein als Künstliche Intelligenz (KI) bezeichnet werden, ihre Spuren auf dem Arbeitsmarkt hinterlassen werden. Welche das genau sein werden, können wir leider aber auch nicht sagen. Denn gerade in solch disruptiven Zeiten, wie wir sie derzeit erleben, wissen wir nicht, wie schnell und in welche Richtung sich bestehende Berufe verändern, welche Berufe verschwinden und welche neu entstehen werden. Zwar können Prognosen etwas darüber sagen, wie sich die Zahl der Berufseinsteiger*innen auf die verschiedenen Berufe und Qualifikationsniveaus verteilen würde, wenn sich die Entwicklung wie in der Vergangenheit fortsetzt. Allerdings scheinen die Potenziale, die sich aus dem Einsatz von KI ergeben, bekannte Zusammenhänge in Frage zu stellen. Hinzu kommt, dass diese Prognosemodelle sehr komplex sind, um daraus sinnvolle Schlussfolgerungen für den Einzelnen zu ziehen. So lässt sich die Frage, inwiefern KI und andere digtale Technolgien auch die Beschäftigungsmöglichkeiten für Menschen mit Behinderungen erweitern könnten, damit kaum beantworten." (Textauszug, IAB-Doku)

Zitatform

Matthes, Britta (2025): Der Einsatz Künstlicher Intelligenz allein kann die zukünftigen Fachkräfteengpässe nicht beheben. (GVG-Perspektive 19), 3 S.

Abstract

"In diesem Beitrag beleuchtet Dr. Britta Matthes, Leiterin der Forschungsgruppe „Berufe in der Transformation“ am IAB, weshalb trotz der Potenziale von Künstlicher Intelligenz Investitionen in die Weiterqualifizierung älterer Beschäftigter unabdingbar bleiben." (Autorenreferat, IAB-Doku)

Zitatform

McGuinness, Seamus, Paul Redmond, Konstantinos Pouliakas, Lorcan Kelly & Luke Brosnan (2025): Technological Change and the Upskilling of European Workers. (IZA discussion paper / Forschungsinstitut zur Zukunft der Arbeit 17753), Bonn, 22 S.

Abstract

"Using the second wave of the European Skills and Jobs survey, this paper measures the relationship between technological change that automates or augments workers' job tasks and their participation in work-related training. We find that 58 per cent of European employees experienced no change in the need to learn new technologies in their jobs during the 2020-21 period. Of those exposed to new digital technology, 14 per cent did not experience any change in job tasks, 10 per cent reported that new tasks had been created while 5 per cent only saw some of their tasks being displaced by new technology. The remaining 13 per cent simultaneously experienced both task displacement and task creation. Our analysis shows that employees in jobs impacted by new digital technologies are more likely to have to react to unpredictable situations, thus demonstrating a positive link between technologically driven task disruption and job complexity. We show a strong linear relationship between technologically driven job task disruption and the need for job-related training, with training requirements increasing the greater the impact of new technologies on task content." (Author's abstract, IAB-Doku) ((en))

Zitatform

Milanez, Anna, Annikka Lemmens & Carla Ruggiu (2025): Algorithmic management in the workplace. New evidence from an OECD employer survey. (OECD Artificial Intelligence Papers 31), Paris, 72 S. DOI:10.1787/287c13c4-en

Abstract

"Algorithmisches Management – der Einsatz von Software, die auch künstliche Intelligenz (KI) umfassen kann, zur vollständigen oder teilweisen Automatisierung von Tätigkeiten, die bislang von menschlichen Führungskräften ausgeführt wurden – hat in den letzten Jahren zunehmend Beachtung gefunden. Einerseits kann algorithmisches Management innerhalb von Unternehmen zu Produktivitäts- und Effizienzgewinnen sowie zu konsistenteren und objektiveren Managemententscheidungen führen. Andererseits gibt es aus anderen Studien immer mehr Anzeichen für seine potenziell negativen Auswirkungen auf die Beschäftigten. Während Politikverantwortliche mit der Frage ringen, wie den Herausforderungen des algorithmischen Managements begegnet werden könnte, bedarf es weiterer Evidenz. Um dieses Ziel zu erreichen, stützt sich der vorliegende Bericht auf eine einzigartige Erhebung unter mehr als 6 000 Unternehmen in sechs Ländern: Deutschland, Frankreich, Italien, Japan, Spanien und die Vereinigten Staaten. Er bietet beispiellose Einblicke in die Prävalenz des algorithmischen Managements, seine wahrgenommenen Auswirkungen und die auf Unternehmensebene ergriffenen Maßnahmen zur Steuerung seines Einsatzes. Die Ergebnisse zeigen, dass algorithmische Management-Systeme in der Mehrzahl der untersuchten Länder bereits weit verbreitet sind. In der Wahrnehmung von Führungskräften verbessert algorithmisches Management zwar häufig die Qualität der Entscheidungen und die Zufriedenheit mit der eigenen beruflichen Situation, die Vertrauenswürdigkeit solcher Systeme gibt ihnen jedoch auch Anlass zur Sorge. Als Gründe dafür nennen sie die unklare Verantwortlichkeit, die nicht leicht nachzuvollziehende Logik der Systeme und den unzureichenden Gesundheitsschutz der Beschäftigten. Es ist dringend erforderlich, Politiklücken zu untersuchen, um den vertrauenswürdigen Einsatz algorithmischer Management-Systeme sicherzustellen." (Autorenreferat, IAB-Doku)

Zitatform

Minniti, Antonio, Klaus Prettner & Francesco Venturini (2025): AI innovation and the labor share in European regions. In: European Economic Review, Jg. 177. DOI:10.1016/j.euroecorev.2025.105043

Abstract

"This paper examines how the development of Artificial Intelligence (AI) affects the distribution of income between capital and labor, and how these shifts contribute to regional income inequality. To investigate this issue, we analyze data from European regions dating back to 2000. We find that for every doubling of regional AI innovation, the labor share declines by 0.5% to 1.6%, potentially reducing it by 0.09 to 0.31 percentage points from an average of 52%, solely due to AI. This new technology has a particularly negative impact on high- and medium-skill workers, primarily through wage compression, while for low-skill workers, employment expansion induced by AI mildly offsets the associated wage decline. The effect of AI is not driven by other factors influencing regional development in Europe or by the concentration of the AI market." (Author's abstract, IAB-Doku, © 2025 The Authors. Published by Elsevier B.V.) ((en))

Zitatform

Muehlemann, Samuel (2025): Artificial intelligence adoption and workplace training. In: Journal of Economic Behavior & Organization, Jg. 238. DOI:10.1016/j.jebo.2025.107206

Abstract

"As artificial intelligence (AI) reshapes business processes, firms must adapt their training strategies to cultivate a skilled workforce. Using German establishment-level panel data from 2019 to 2023, this study analyzes how firms adjust their training strategies following AI adoption. Staggered difference-in-differences analysis shows that sustained AI adoption is associated with a 14% increase in new apprenticeships among training firms (intensive margin), but is not linked to the training decision (extensive margin). AI adoption is also associated with a modest increase in continuing training, with resources shifting toward high-skilled employees. The results align with AI as an automation innovation that reduces demand for simple skills as well as an augmentation innovation that increases demand for more advanced skills. The German dual apprenticeship system appears critical for firms aiming to build a future-ready workforce in the age of AI." (Author's abstract, IAB-Doku, © 2025 The Author. Published by Elsevier B.V.) ((en))

Zitatform

Müller, Steffen & Verena Plümpe (2025): Wie Roboter die betriebliche Beschäftigungsstruktur verändern. In: Wirtschaft im Wandel, Jg. 31, H. 1, S. 10-13. DOI:10.18717/wwfyns-ep75

Abstract

"Der Einsatz von Robotern verändert die Arbeitswelt grundlegend – doch welche spezifischen Effekte hat dies auf die Beschäftigungsstruktur? Unsere Analyse untersucht die Folgen des Robotereinsatzes anhand neuartiger Mikrodaten aus deutschen Industriebetrieben. Diese Daten verknüpfen Informationen zum Robotereinsatz mit Sozialversicherungsdaten und detaillierten Angaben zu Arbeitsaufgaben. Auf Basis eines theoretischen Modells leiten wir insbesondere positive Beschäftigungseffekte für Berufe mit wenig repetitiven, programmierbaren Aufgaben ab, sowie für jüngere Arbeitskräfte, weil diese sich besser an technologische Veränderungen anpassen können. Die empirische, mikroökonomische Analyse des Robotereinsatzes auf Betriebsebene bestätigt diese Vorhersagen: Die Beschäftigung steigt für Techniker, Ingenieure und Manager und junge Beschäftigte, während sie bei geringqualifizierten Routineberufen sowie bei Älteren stagniert. Zudem steigt die Fluktuation bei geringqualifizierten Arbeitskräften signifikant an. Unsere Ergebnisse verdeutlichen, dass der Verdrängungseffekt von Robotern berufsabhängig ist, während junge Arbeitskräfte neue Tätigkeiten übernehmen." (Autorenreferat, IAB-Doku)

Zitatform

Müller, Christoph (2025): Weiterbildungsungleichheit und technologischer Wandel: Nach IT-Investitionen steigt vor allem die Weiterbildungsquote der Höherqualifizierten. (IAB-Kurzbericht 06/2025), Nürnberg, 8 S. DOI:10.48720/IAB.KB.2506

Abstract

"Betriebliche Weiterbildung soll dazu beitragen, die Fähigkeiten der Beschäftigten an neue Anforderungen anzupassen. Gerade im Zuge der digitalen Transformation der Arbeitswelt sind solche Anpassungen dringend erforderlich. Die vorliegende Analyse des Zusammenhangs zwischen Investitionen in digitale Technologien und innerbetrieblicher Weiterbildung zeigt: In Betrieben mit IT-Investitionen steigen die Weiterbildungsquoten der Beschäftigten mit qualifizierten Tätigkeiten; bei denjenigen mit einfachen Tätigkeiten ist dies im Mittel hingegen nicht der Fall." (Autorenreferat, IAB-Doku)

Zitatform

Neumann, Uwe (2025): Digital transformation, employment change and the adaptation of regions in Germany. In: Structural Change and Economic Dynamics, Jg. 73, S. 37-50. DOI:10.1016/j.strueco.2024.12.014

Abstract

"Digital change is often said to lead to large-scale job losses. Using data from administrative sources in Germany, this study examines the extent to which adaptation to digital change has affected regional employment growth and disparities over the past decade. The analysis confirms previous research according to which increases in productivity coincide with regional job growth rather than decline. Incorporating various indicators of digitalisation and automation into a model of industry-specific regional job growth shows that local labour markets with very different characteristics – regions with strong manufacturing clusters on the one hand and large cities on the other – have achieved employment growth despite high automation exposure. While the study highlights regional differentials with respect to the adaptation to technological change, less prosperous regions may face a much greater challenge in realising job creation potentials. The results argue against policy efforts aimed at “protecting” jobs from digitalisation and automation." (Author's abstract, IAB-Doku, © 2025 The Author. Published by Elsevier B.V.) ((en))

Zitatform

Nigar, Meher, Jannatul Ferdous Juli, Uttam Golder, Mohammad Jahangir Alam & Mohammad Kamal Hossain (2025): Artificial intelligence and technological unemployment. Understanding trends, technology's adverse roles, and current mitigation guidelines. In: Journal of open innovation, Jg. 11, H. 3. DOI:10.1016/j.joitmc.2025.100607

Abstract

"As artificial intelligence (AI) and automation continue to reshape industries, concerns about technological unemployment are intensifying. This study employs a Systematic Literature Review (SLR) guided by the PRISMA framework to examine peer-reviewed literature from the Scopus database (2015–July 09, 2025). It identifies threecore themes: (1) trends in AI-induced labor displacement, including task automation, skill polarization, and industry-specific disruptions in sectors such as healthcare, education, and creative industries; (2) the adverse roles of AI technologies, particularly in affecting white-collar professionals, gig workers, and freelancers by increasing precarity and skill mismatches; and (3) existing mitigation strategies, including responsible AI guidelines proposed by governments, institutions, and firms aimed at balancing technological advancement with employment protection. While a growing body of policy responses encourages human-AI complementarity, current measures remain fragmented and insufficient to address the structural risks of workforce displacement. This study presents a comprehensive synthesis of the evolving relationship between AI and employment, highlighting key areas for further inquiry and policy development." (Author's abstract, IAB-Doku, © 2025 The Authors. Published by ElsevierLtd on behalf of Prof JinHyo Joseph Yun.) ((en))

Zitatform

Nikolova, Milena, Anthony Lepinteur & Femke Cnossen (2025): Just another cog in the machine? A worker‐level view of robotization and tasks. In: Economica, Jg. 92, H. 368, S. 1101-1148. DOI:10.1111/ecca.70006

Abstract

"Technological change has led to a decline in the share of routine and physical jobs, and a rise in the share of abstract and social ones at the economy level. However, much less is known about how these trends unfold at the individual level. Do workers' tasks become more or less routine and physical? Do workers shift towards more social and abstract activities? This paper is the first to explore these questions in the context of robotization. We use survey data from 20 European countries to develop worker-level indices of physical, routine, abstract and social tasks, which we link to industry-level robotization exposure. Using instrumental variable techniques, we find that robotization reduces physically demanding tasks but increases routine tasks, while also limiting opportunities for cognitively challenging work and human interaction. This study provides a worker-centric perspective on the relationship between technology and task composition, revealing insights that aggregate analyses miss." (Author's abstract, IAB-Doku) ((en))

Zitatform

Ohly, Sandra, Amina Granica, Hannes Herwig, Niklas Kühlewind, Tobias Paetzold, Lana Vinhoven, Luisa Wiemers & Xi Zhang (2025): Humane work design: How does the digitization of work change core work characteristics in different occupations? In: Zeitschrift für Arbeitswissenschaft, Jg. 79, H. 4, S. 569-580. DOI:10.1007/s41449-025-00491-4

Abstract

"Wenn vermehrt digitale Tools eingesetzt werden und Arbeit im Homeoffice stattfindet, verändern sich Arbeitsbedingungen. Der vorliegende Beitrag analysiert diese Veränderungen anhand von 8 Fällen und arbeitet Gemeinsamkeiten und Unterschiede heraus. Die Prognose von Lernen, Wohlbefinden und Arbeitsleistung erlaubt es, Fehlentwicklungen in der Arbeitsgestaltungen abzuschätzen.Praktische Relevanz: Durch Aufmerksamkeit auf mögliche Fehlentwicklungen können Gegenmaßnahmen getroffen werden, etwa zur Verbesserung der Kommunikation und sozialen Unterstützung bei ortsverteilter Arbeit, oder der angepassten Gestaltung von KI-Systemen." (Autorenreferat, IAB-Doku)

Zitatform

Okur, Fatih & Enes Özdemir (2025): Artificial intelligence, automation and employment dynamics: empirical evidence from G7 economies. In: Journal of Economic Studies, S. 1-17. DOI:10.1108/jes-06-2025-0414

Abstract

"Purpose: This study examines how the rapid adoption of artificial intelligence (AI) and automation affects employment dynamics across G7 economies. While previous research has often focused on either AI or robotics in isolation, their combined and long-term effects on employment remain poorly understood. Addressing this gap is crucial for policymakers seeking to balance technological progress with labor market stability. Design/methodology/approach: Using a balanced panel dataset covering 2010–2024 for the G7 countries, thestudy investigates the relationships between AI investment (proxied by information and communication technology (ICT) investment), robot density (ROBOT), wages, productivity (PRD) and education spending (EDU), and their impact on employment. The analysis employs panel unit root and cross-sectional dependence tests, a panel autoregressive distributed lag (ARDL) framework estimated via the pooled mean group (PMG) estimator, and robustness checks using Driscoll–Kraay fixed effects, common correlatedeffects (CCE) estimators, country-specific regressions and Dumitrescu –Hurlin panel causality tests. Findings: The results reveal that AI investment has a significant negative effect on employment in the long run, whereas ROBOT shows a positive but context-dependent relationship. Wage levels are negatively associated with employment, while PRD shows only a modest positive influence. Education expenditure exhibits mixed behavior – positive in the short run but negative in the long run – suggesting potential misalignment with evolving labor market needs. Causality tests confirm a unidirectional link from AI investment to employment, underscoring its structural role in labor market change. Research limitations/implications: This study is limited by data availability, particularly the lack of detailed sectoral or occupational breakdowns across countries. As a result, it cannot fully capture the distributional effects of AI and automation across different worker groups. The use of proxies, such as ICT investment for AI, may not reflect the full scope of AI deployment. Despite these limitations, the findings highlight important macro-level dynamics and suggest that technological investments significantly shape employment trends. Future research should utilize micro-level data to explore sector-specific impacts, wage effects and labor force transitions in response to digital transformation. Practical implications: The findings suggest that without targeted policy interventions, increased AI investment may displace workers in the long run. Policymakers should prioritize reskilling, adapt education systems to evolving technological needs, and differentiate strategies across sectors and worker skill levels. Social implications: This study highlights the potential for AI and automation to reshape labor markets, with implications for income distribution, job security and social cohesion. The displacement of routine jobs may disproportionately affect low-skilled and vulnerable workers, increasing the risk of inequality and social exclusion. To prevent deepening divides, social policies must focus on equitable access to education, digital literacy and lifelong learning. Supporting workforce adaptability through inclusive training programs and social safety nets is essential. The results underscore the urgent need for collaborative efforts between governments, educational institutions and industries to ensure a socially sustainable digital transformation. Originality/value: This study is among the first to jointly analyze AI and robotics within a dynamic panel framework, offering new cross-country evidence on their heterogeneous employment effects in advanced economies. By integrating multiple estimation strategies and country-specific perspectives, the paper contributes to a more nuanced understanding of how technological transformation reshapes labor markets and highlights the institutional conditions that mediate these effects." (Author's abstract, IAB-Doku, © EmeraldGroup) ((en))

Zitatform

Otoiu, Adrian, Emilia Titan, Dorel Paraschiv & Daniela-Ioana Manea (2025): Job polarisation OR AND upgrading! Recent evidence from Europe. In: The Economic and Labour Relations Review, Jg. 36, H. 1, S. 257-270. DOI:10.1017/elr.2025.12

Abstract

"Based on recent evidence from Europe, the paper shows that polarization and upgrading are not mutually exclusive trends, but rather, simultaneously defined recent structural changes in employment. The results show that (a) the occupational structure shows a general shift towards high-skill jobs, (b) the prevailing upgrading patterns are often accompanied by job polarization, as the share of middle-skill jobs declines in most cases, and (c) while low-skill employment often outperforms middle-skill jobs, it has tended to decline. In addition to analysing trends for EU-27 countries with different levels of development for the latest available time periods, the article also shows that occupational upgrading patterns are rather intertwined with job polarisation and are compatible with both the Skill-Biased Technical Change (SBTC) and Routine-Biaszd Technical Change (RBTC) hypotheses. The employment dynamics of low-skill workers are uncertain, as they are not fully compatible with any theoretical model, thus pointing to the need for a finer understanding of changes in occupational structure, and the extent to which both polarization and upgrading are shaping the evolution of the labor force structure under the impact of (ongoing) technological change." (Author's abstract, IAB-Doku) ((en))

Zitatform

Ottaiano, Mario Michael, Lea Schneidemesser & Florian Butollo (2025): Künstliche Intelligenz und Erfahrungswissen. Zur Formalisierbarkeit und Delokalisierung von Facharbeit. In: Arbeit. Zeitschrift für Arbeitsforschung, Arbeitsgestaltung und Arbeitspolitik, Jg. 34, H. 4, S. 275-298. DOI:10.1515/arbeit-2025-0015

Abstract

"Der Beitrag untersucht aus arbeitssoziologischer Perspektive, wie der Einsatz von KI-Systemen in der prädiktiven Instandhaltung (PdM) die Bedeutung von Erfahrungswissen im Arbeitsprozess verändert und wie sich die Zusammenarbeit zwischen KI-Anwender- und KI-Anbieterunternehmen gestaltet. Die Fallstudie in einem Unternehmen der Papierindustrie zeigt, dass die Grenze der Formalisierbarkeit von Erfahrungswissen in der Instandhaltungsarbeit graduell ausgeweitet werden kann, wobei das domänenspezifische Wissen der Beschäftigten weiterhin eine wichtige Rolle für die Verbesserung der Software und die Interpretation der gewonnenen Daten spielt. Weil die Einführung des KI-Systems auch eine Delokalisierung von Wissen ermöglicht, kommt es zugleich zu einer konfliktträchtigen Verschiebung der Kompetenzen vom Anwenderunternehmen zum Softwareanbieter." (Autorenreferat, IAB-Doku)

Zitatform

Parent-Thirion, Agnes, Nora Wukovits-Votzi & Jessye Muller (2025): Future-oriented occupations in the EU. Main features, employment conditions, and job strain. 51 S. DOI:10.2767/2953537

Abstract

"The way we work is changing due to developments associated with the digital and green transition as well as demographic change, as a driver of current and future labour shortages. As these transitions impact job content, tasks and processes, they will change how people work, the skills needed to carry out jobs, employment conditions, and, ultimately, dimensions of their job quality. These transition-related changes in occupations are of high relevance for workers, job applicants, and students training to join these occupations, as well as stakeholders, and policy makers, at the sectoral, national, and European levels. While their impacts are separately treated in this analysis, the green and digital transitions can further exacerbate labour shortages given the skill profiles required by related occupations." (Text excerpt, IAB-Doku) ((en))

Zitatform

Peeters, Maria C. W., Jan Fekke Ybema, Pascale M. Le Blanc & Judith Plomp (2025): Exploring the delicate relation between technological innovations and work quality: A study among civil servants. In: Economic and Industrial Democracy, Jg. 46, H. 3, S. 851-873. DOI:10.1177/0143831x251347151

Abstract

"This study explores the delicate relation between technological innovations and work quality. It was conducted across various parts of the Dutch central government. The authors assessed how civil servants perceive changes in job demands, job resources and some relevant outcomes following the implementation of new technologies. Data were collected through an online Technology Monitor (TM) which was (at least partly) completed by 332 respondents. Results showed that employees perceived significant increases in various job demands, alongside a modest increase in the job resource autonomy after technology implementation. Additionally, civil servants who experienced more autonomy following new technology implementation reported higher levels of both work engagement and employability. In contrast, perceptions of increased workload were associated with more burnout symptoms. Interestingly, perceived increases in task variation were associated with fewer burnout symptoms, lower job insecurity and higher work engagement. These findings offer valuable insights for managers and HR professionals involved in managing technological transitions, emphasizing the importance of employee-centered strategies to safeguard and enhance the quality of work of civil servants." (Author's abstract, IAB-Doku) ((en))

Zitatform

Pouliakas, Konstantinos & Giulia Santangelo (2025): Are Artificial Intelligence (AI) Skills a Reward or a Gamble? Deconstructing the AI Wage Premium in Europe. (IZA discussion paper / Forschungsinstitut zur Zukunft der Arbeit 17607), Bonn, 33 S.

Abstract

"Understanding the labor market impact of new, autonomous digital technologies, particularly generative or other forms of artificial intelligence (AI), is currently at the top of the research and policy agenda. Many initial studies, though not all, have shown that there is a wage premium to AI skills in labor markets. Such evidence tends to draw on data from web-based sources and typically deploys a keyword approach for identifying AI skills. This paper utilizes representative adult workforce data from 29 European countries, the second European skills and jobs survey, to examine wage differentials of the AI developer workforce. The latter is uniquely identified as part of the workforce that writes programs using AI algorithms. The analysis shows that, on average, AI developers enjoy a significant wage premium relative to a comparably educated or skilled workforce, such as programmers who do not yet write code using AI at work. Wage decomposition analysis further illustrates that there is a large unexplained component of such wage differential. Part of AI programmers' larger wage variability can be attributed to a greater performance-based component in their wage schedules and higher job-skill requirements." (Author's abstract, IAB-Doku) ((en))

Zitatform

Pouliakas, Konstantinos, Giulia Santangelo & Paul Dupire (2025): Are artificial intelligence skills a reward or a gamble? Deconstructing the AI wage premium in Europe. In: Eurasian business review, Jg. 15, H. 4, S. 1091-1128. DOI:10.1007/s40821-025-00302-0

Abstract

"Understanding the labor market impact of new, autonomous digital technologies, particularly generative or other forms of artificial intelligence (AI), is currently at the top of the research and policy agenda. Many initial studies, though not all, have shown that there is a wage premium to mostly technical AI skills in labor markets. Such evidence tends to draw on data from web-based sources and typically fails to provide insight into the mechanisms underlying the AI wage gap. This paper utilizes representative adult workforce data from 29 European countries, the second European skills and jobs survey, to examine wage differentials of the AI programmer workforce. The latter is uniquely identified as part of the workforce that writes computer programs using AI algorithms. The analysis shows that, on average, AI programmers enjoy a significant wage premium relative to a comparably educated or skilled workforce, such as programmers who do not yet write code using AI at work. Wage decomposition analysis further illustrates that there is a large unexplained component of such wage differential. Part of AI programmers’ larger wage variability can however be attributed to higher job-skill requirements, a propensity for remote work and a greater performance-based component in wage schedules. This indicates differences in the job design and performance management of the AI workforce." (Author's abstract, IAB-Doku) ((en))

Zitatform

Rabensteiner, Thomas & Alexander Guschanski (2025): Occupational Autonomy and Wage Divergence: Evidence From European Survey Data. In: BJIR, Jg. 63, H. 4, S. 696-713. DOI:10.1111/bjir.70003

Abstract

"Wages across occupations in Western Europe have diverged, resulting in increased wage inequality. However, existing theories such as routine-biased technological change (RBTC) or task offshoring fail to explain this trend. We propose a new explanation based on occupational autonomy. Autonomy measures workers' control and influence over their work process based on the tasks required in an occupation. Analysing individual-level data from the EU Survey of Income and Living Conditions, we reveal a rising autonomy wage premium, that is, higher wage growth for occupations with higher autonomy, which accurately predicts the observed occupational wage divergence. We also find that the autonomy premium increases more rapidly in countries and industries with greater employee monitoring and outsourcing, as well as in countries with declining minimum wages. These findings imply that low-autonomy occupations have been disadvantaged by recent socioeconomic trends that have altered power relations in the workplace. Notably, our analysis does not support previous explanations for occupational wage trends based on RBTC or task offshoring." (Author's abstract, IAB-Doku) ((en))

Zitatform

Rehmer, Sabine, Holger Muehlan, Maren Menzel, Maike Juds, Michael Fellmann, Hitesh Dhiman & Carsten Röcker (2025): Die Hybridisierung von menschlichen und technischen Arbeitsleistungen mit Künstlicher Intelligenz als neuer Leittechnologie: Entwicklungen, Implikationen und Potenziale für menschengerechte Arbeit. In: Zeitschrift für Arbeitswissenschaft, Jg. 79, H. 4, S. 525-533. DOI:10.1007/s41449-025-00489-y

Abstract

"Der Artikel untersucht die Hybridisierung menschlicher und technischer Arbeitsleistungen im Kontext Künstlicher Intelligenz (KI) als neuer Leittechnologie und analysiert ihre Entwicklungen, Implikationen und Potenziale für eine menschengerechte Arbeitsgestaltung. Ausgehend von der Beobachtung, dass KI zunehmend kognitive Tätigkeiten transformiert, wird mit dem Konzept des „Hybrid Man“ ein neues Menschenbild skizziert, das die enge Verschränkung von menschlicher und maschineller Intelligenz beschreibt. Damit einher gehen tiefgreifende Herausforderungen: Die Grenzen zwischen menschlicher und technischer Leistung verschwimmen, Verantwortungsfragen und rechtliche Unsicherheiten entstehen, während zugleich neue Anforderungen an Kompetenzen wie „AI Literacy“ sichtbar werden. Ebenso rücken Fragen nach Transparenz, Erklärbarkeit und Akzeptanz von KI-Systemen in den Vordergrund, die für Vertrauen und nachhaltige Integration entscheidend sind. Neben den Chancen zur Entlastung und Erweiterung menschlicher Fähigkeiten birgt die Hybridisierung Risiken wie Überwachungsdruck, Anpassungsstress und negative psychische Beanspruchungen. Vor diesem Hintergrund schlägt der Artikel mit dem Konzept des eudaimonischen Wohlbefindens („Eudaimonia“) ein normatives Kriterium für die Gestaltung zukünftiger Arbeit vor, das Wachstum, Sinnhaftigkeit, Authentizität und Exzellenz fördert. Ziel ist eine interdisziplinär fundierte, menschengerechte Arbeitswelt, die die Potenziale von KI nutzbar macht, ohne die psychische Gesundheit, Selbstwirksamkeit und Würde der Beschäftigten zu gefährden." (Autorenreferat, IAB-Doku)

Zitatform

Ribeiro, Marcos J. & Klaus Prettner (2025): The Skill Premium Across Countries in the Era of Industrial Robots and Generative AI. (Department of Economics working paper / Vienna University of Economics and Business 381), Wien, 29 S.

Abstract

"How do new technologies affect economic growth and the skill premium? To answer this question, we analyze the impact of industrial robots and artificial intelligence (AI) on the wage differential between low-skill and high-skill workers across 52 countries using counterfactual simulations. In so doing, we extend the nested CES production function framework of Bloom et al. (2025) to account for cross-country income heterogeneity. Confirming prior findings, we Show that the use of industrial robots tends to increase wage inequality, while the use of AI tends to reduce it. Our contribution lies in documenting substantial heterogeneity across income groups: the inequality-increasing effect of robots and the inequality-reducing effects of AI are particularly strong in high-income countries, while they are less pronounced among middle- and lower-middle income countries. In addition, we show that both technologies boost economic growth. In terms of policy recommendations, our findings suggest that investments in education and skill-upgrading can simultaneously raise average incomes and mitigate the negative effects of automation on wage inequality." (Author's abstract, IAB-Doku) ((en))

Zitatform

Riccio, Federico, Jacopo Staccioli & Maria Enrica Virgillito (2025): European regional employment and exposure to labour-saving technical change: results from a direct text similarity measure. (LEM working paper series / Laboratory of Economics and Management 2025/19), Pisa, 34 S. DOI:10.57838/sssa/02jp-b197

Abstract

"Does labor-saving technological change pose a threat to European employment, and if so, to what extent? This study investigates the degree of employment exposure to labor-saving technological change across NUTS-2 regions in Europe. We construct a cross-walked metric between the SOC and ISCO classification systems to adapt the direct measure of occupational exposure developed by Montobbio et al. (2024) for the US economy and apply it to the European context. This methodology enables us to generate detailed insights into the exposure of European occupations by leveraging the similarity rankings between technological classifications in the USPTO (CPCs) and task descriptions. To evaluate the transmission from occupational exposure to employment outcomes, we utilise data from the European Structure of Earnings Survey (EU-SES), thereby constructing exposure indices at both sectoral and regional levels. Finally, we examine the industrial and geographical diffusion of labor-saving technological change in recent years and provide robust econometric evidence indicating that low-wage regions, as well as deindustrialising areas heavily integrated into global value chains, are disproportionately vulnerable to the threat of substitution." (Author's abstract, IAB-Doku) ((en))

Zitatform

Richiardi, Matteo Guido, Leonie Westhoff, Caterina Astarita, Ekkehard Ernst, Clare Fenwick, Neysan Khabirpour & Lorenzo Pelizzari (2025): The impact of a decade of digital transformation on employment, wages, and inequality in the EU: a “conveyor belt” hypothesis. In: Socio-economic review, Jg. 23, H. 3, S. 1225-1251. DOI:10.1093/ser/mwaf011

Abstract

"We study the effects of digital transformation in the European Union on individual employment outcomes, wage growth, and income inequality, during the decade 2010–9. Our results allow us to formulate a ‘conveyor-belt’ hypothesis suggesting that employment confers a competitive advantage in navigating the digital transition due to the accumulation of pertinent skills in the workplace. Because digital skills are acquired with the changing demands of the job, their initial endowment matters less for the employed than for the non-employed. Furthermore, the ability of out-of-work individuals with higher digital skills to jump back on the labour market is reduced for those with higher education, suggesting a faster depreciation of their digital skills. A similar effect, although of limited size, is found for earning growth: out-of-work individuals with higher digital skills are not only more likely to find a job, but experience higher earnings growth, compared to their peers with lower digital skills. Our results point to a vulnerability of workers ‘left behind’ from the digital transformation and the labour market. The overall effects on inequality are, however, limited." (Author's abstract, IAB-Doku) ((en))

Zitatform

Sartori, Laura & Clementine Collett (2025): Sociotechnical imaginaries of social inequality in the design and use of AI recruitment technology. In: European Societies, Jg. 27, H. 3, S. 409-432. DOI:10.1162/euso_a_00035

Abstract

"Through interviewing 12 companies in Italy which either design (vendors) or use (clients) AI recruitment technology systems, we explore how these companies perceive their systems to interact with issues of social inequality and how these perceptions, in practice, carry societal impacts. Three sociotechnical imaginaries (Jasanoff and Kim, 2015) were consistently embedded within these companies’ visions of this intersection: the third eye, the river, and the car bonnet. Through critically analyzing these imaginaries, we find that they exhibit an overriding desire for productivity and talent capture from clients, and a consequential de-prioritization of addressing social inequality and scrutinizing the ways it could be reproduced from both vendors and clients. It demonstrates that the current ‘desired’ futures, shown by the sociotechnical imaginaries which vendors and clients share for AI-tec-tech are really leading us towards an ‘undesirable’ future of hiring which continues to perpetuate social inequality. This study contributes one of the first pieces of empirical work to simultaneously assess the perceptions of AI-rec-tech vendors ’ and clients’ surrounding social inequality, to shed light on the priorities for design and the motivations for usage, and to reflect upon how this impacts society. This is a significant and original contribution to the evolving body of literature on AI-rec-tech in sociology, critical data studies, and communications." (Author's abstract, IAB-Doku) ((en))

Zitatform

Schaal, Jacob (2025): A theory-based AI automation exposure index: Applying Moravec's Paradox to the US labor market. (arXiv papers), 33 S. DOI:10.48550/arXiv.2510.13369

Abstract

"This paper develops a theory-driven automation exposure index based on Moravec's Paradox. Scoring 19,000 O*NET tasks on performance variance, tacit knowledge, data abundance, and algorithmic gaps reveals that management, STEM, and sciences occupations show the highest exposure. In contrast, maintenance, agriculture, and construction show the lowest. The positive relationship between wages and exposure challenges the notion of skill-biased technological change if AI substitutes for workers. At the same time, tacit knowledge exhibits a positive relationship with wages consistent with seniority-biased technological change. This index identifies fundamental automatability rather than current capabilities, while also validating the AI annotation method pioneered by Eloundou et al. (2024) with a correlation of 0.72. The non-positive relationship with pre-LLM indices suggests a paradigm shift in automation patterns." (Author's abstract, IAB-Doku) ((en))

Zitatform

Schneemann, Christian, Florian Bernardt, Michael Kalinowski, Tobias Maier, Gerd Zika & Marc Ingo Wolter (2025): Auswirkungen des Strukturwandels auf die Arbeitsmarktregionen und Bundesländer in der langen Frist – Qualifikations- und Berufsprojektion bis 2040. (IAB-Forschungsbericht 03/2025), Nürnberg, 46 S. DOI:10.48720/IAB.FB.2503

Abstract

"Die Bundesländer und die Arbeitsmarktregionen in Deutschland unterscheiden sich in ihrer Bevölkerungs- und Wirtschaftsstruktur, weshalb sie auch unterschiedliche Arbeitskräfteengpässe und -überhänge aufweisen. Aufgrund ihrer verschiedenartigen Entwicklungen werden auch künftig Unterschiede im Arbeitsmarktgeschehen bestehen. Mit Hilfe des sogenannten QuBe-Modellverbundes (8. Welle der QuBe-Basisprojektion) werden langfristige immanente Megatrends wie die demografische Entwicklung, der wirtschaftliche Strukturwandel und die Digitalisierung im Modell selbst erfasst und die Auswirkungen auf Wirtschaft und Arbeitsmarkt sichtbar gemacht. Die Analyse zeigt, dass sich die wirtschaftliche Lage in Deutschland nicht wie in der Vergangenheit durch positive Entwicklungen im Außenhandel erholen wird. Das zukünftige Handeln der USA, China und Russlands ist schwer abzuschätzen und erhöht die Unsicherheit auf dem Weltmarkt. Zudem wird das künftige Arbeitsmarktgeschehen zu einem großen Teil von der demografischen Entwicklung, dem stetigen strukturellen Wandel (z.B. Digitalisierung im Handel) und der schwächeren Nachfrage im Baugewerbe geprägt. So wird das Arbeitskräfteangebot infolge des Bevölkerungsrückgangs in vielen Bundesländern und Arbeitsmarktregionen bis zum Jahr 2040 sinken. Zwar können einige Arbeitsmarktregionen noch Bevölkerung aufbauen, aber die Bevölkerung im erwerbsfähigen Alter wird in allen abnehmen. Infolgedessen wird auch der Arbeitskräftebedarf fast überall sinken. Insgesamt wird in vielen Bundesländern und Arbeitsmarktregionen die Erwerbslosenquote sinken oder nahezu stabil bleiben, so dass dort trotz der schlechteren konjunkturellen Entwicklung weiterhin mit Engpässen in verschiedenen Wirtschaftsbereichen und Berufen zu rechnen ist. Die Rekrutierung von Arbeitskräften dürfte somit in vielen Wirtschaftsbereichen und Regionen langfristig zunehmend schwieriger werden. Der Bedarf an qualifiziertem Personal im Wirtschaftszweig „Heime und Sozialwesen“ oder im Bereich der IT-Dienstleistungen wächst kontinuierlich. Dies alles geschieht vor dem Hintergrund eines wohl eher noch beschleunigten Strukturwandels, der gerade die Bundesländer und Arbeitsmarktregionen schon jetzt vor große Herausforderungen stellt, in denen das Verarbeitende Gewerbe zum Beispiel die Automobilindustrie nach wie vor überdurchschnittlich. Die fortschreitende Digitalisierung und Dekarbonisierung erfordern eine permanente Modernisierung und Innovationsfähigkeit der deutschen Wirtschaft. Gerade die ökologische Transformation ist stark auf Erwerbstätige im Baugewerbe angewiesen. Eine Qualifizierung in diesem Bereich bleibt deshalb wichtig, weil die Rekrutierungssituation für Unternehmen trotz der langfristig vermutlich zurückgehenden Erwerbstätigkeit im Vergleich zu anderen Berufen auch in Zukunft schwierig sein wird. Die Umsetzung zusätzlich notwendiger Investitionen im Zuge dieser Transformation sollte nicht an fehlenden Fachkräften scheitern." (Autorenreferat, IAB-Doku)

Zitatform

Seele, Stefanie & Oliver Stettes (2025): KI-Nähe im Job zahlt sich aus. (IW-Kurzberichte / Institut der Deutschen Wirtschaft Köln 2025,45), Köln, 3 S.

Abstract

"Beschäftigte, deren Aufgaben eine Nähe zu den Anwendungspotenzialen von Künstlicher Intelligenz (KI) aufweisen, erhalten höhere Tagesentgelte als Beschäftigte in KI-fernen Tätigkeiten. Sie wechseln zudem seltener den Betrieb und haben seltener längere Arbeitslosigkeitsperioden. Die Sorge vor einer zunehmenden Verbreitung von KI im Arbeitsalltag scheint bisher unbegründet." (Autorenreferat, IAB-Doku)

Zitatform

Storm, Eduard, Myrielle Gonschor & Marc Justin Schmidt (2025): AI in Demand: How Expertise Shapes its (Early) Impact on Workers. (Ruhr economic papers 1185), Essen, 46 S. DOI:10.4419/96973370

Abstract

"We study how artificial intelligence (AI) affects workers' earnings and employment stability, combining German job vacancy data with administrative records from 2017-2023. Identification comes from changes in workers' exposure to local AI skill demand over time, instrumented with national demand trends. We find no meaningful displacement or productivity effects on average, but notable skill heterogeneity: expert workers with deep domain knowledge gain while non-experts often lose, with returns shaped by occupational task structures. We also document AI-driven reinstatement effects toward analytic and interactive tasks that raise earnings. Overall, our results imply distributional concerns but also job-augmenting potential of early AI technologies." (Author's abstract, IAB-Doku) ((en))

Zitatform

Stöckl, Andreas & Olaf Struck (2025): Continuous vocational education and training and new technologies: on the importance of educational level and technology in the workplace. In: Journal for labour market research, Jg. 59. DOI:10.1186/s12651-025-00398-x

Abstract

"Continuing vocational education and training (CVET) can support technical and digital developments. At the same time, company-based training increases employability in the face of technical rationalization. Multiple studies show that the highly qualified undertake further training more often than the less qualified. This increases their educational and employment inequality. However, it is unclear whether this general finding also applies to participation in CVET when new technologies are introduced in companies. Companies could have an interest in the deployment of all their employees. The introduction of new technologies in a company context could lead to people with professional qualifications as well as those with higher qualifications taking part in CVET. Using the Linked Personnel Panel (LPP) and the IAB Establishment Panel, we investigate whether the participation of employees with university-level education and upper secondary vocational education differ if the participation in CVET is attributed directly to the introduction of technology in the workplace. The outcomes show that the more highly qualified use forms of self-study more frequently. However, there are only marginal differences between highly qualified and professionally qualified employees when it comes to the use of courses as part of in-house training." (Author's abstract, IAB-Doku) ((en))

Zitatform

Sztandar-Sztanderska, Karolina (2025): How context matters: Human oversight of automated decision-making systems in welfare administration. In: Journal of European Social Policy, S. 1-15. DOI:10.1177/09589287251358069

Abstract

"This article examines the capacity of frontline staff to oversee automated decision-making (ADM) systems, which are increasingly used in digital welfare states to make life-altering decisions. While prior research had focused on cognitive limitations that lead to human over-reliance on ADM, the role of contextual factors that also shape caseworkers’ supervisory activities has neither been conceptualized, nor systematically investigated. To address this gap, we develop an analytical framework for the context-sensitive study of frontline oversight, inspired by the street-level bureaucracy perspective and human–computer interaction studies. We also demonstrate the framework’s relevance through findings from a mixed-methods study of a profiling algorithm used by Public Employment Services in Poland. We identify four types of factors – policy-, organization-, professionalism-, and technology-related – that effectively shape frontline oversight. Our findings also have practical implications, as the inclusion of humans in the decision-making loop is a central element of regulatory efforts aimed at protecting individuals from algorithmic harms." (Author's abstract, IAB-Doku) ((en))

Zitatform

Teutloff, Ole & Fabian Braesemann (2025): Resilienz statt Reskilling: Wie KI den Arbeitsmarkt verändert und wie wir darauf reagieren müssen. Teil des Zeitgesprächs: "Beschäftigung im Wandel: Wie KI, Demografie und Institutionen den Arbeitsmarkt verändern". In: Wirtschaftsdienst, Jg. 105, H. 10, S. 715-719. DOI:10.2478/wd-2025-0184

Abstract

"Generative künstliche Intelligenz (KI) verändert die Arbeitsmärkte, indem sie Aufgaben und Berufe, die Nachfrage nach Qualifikationen und die Machtverhältnisse zwischen Arbeitnehmern und Arbeitgebern verändert. Im Gegensatz zu früheren technologischen Umbrüchen betrifft KI auch Hochqualifizierte. Doch wie müssen wir darauf reagieren, wenn KI immer mehr Tätigkeiten übernehmen kann? Statt kurzfristiger Umschulungsprogramme braucht es langfristige Strategien, die die Widerstandsfähigkeit von Beschäftigten und Strukturen stärken – und so eine nachhaltige Anpassung des Arbeitsmarktes an die sich wandelnde technologische Landschaft ermöglichen." (Autorenreferat, IAB-Doku)

Zitatform

Vasiakina, Mariia & Christian Dudel (2025): The hidden costs of technological change: investigating pathways through which highly automatable jobs undermine workers’ health in Germany. (MPIDR working paper / Max Planck Institute for Demographic Research 2025-032), Rostock, 29 S. DOI:10.4054/mpidr-wp-2025-032

Abstract

"The ongoing economic transformation driven by automation has significant social implications, particularly for the health and well-being of workers who face the risk of job displacement and the pressure to acquire new skills and qualifications. However, the specific pathways through which exposure to automation risk affects health outcomes remain poorly understood, and the relative contribution of each potential mechanism is still unclear. In this study, we examine the nature of the relationship between high workplace exposure to automation risk and a range of subjective health outcomes – including self-reported health, anxiety, and both physical and mental component summary scores from the SF-12 Health Survey – among workers in Germany. Using data from the German Socio-EconomicPanel (SOEP) linked with administrative records from the Occupational Panel for Germany (2014–2022), we apply the Karlson-Holm-Breen (KHB) mediation analysis method to assess whether broader indicators of economic uncertainty, alongside automation-specific factors, mediate the relationship between high automation risk and workers’ health. Our results indicate that the negative impact of high automation risk on health in Germany primarily operates through indirect pathways (related to mediators) for both genders, with the exception of physical health among male workers, where a direct negative effect is also evident. Economic concerns – particularly job insecurity and worries about one’s future financial situation – emerge as more significant mediators than automation-specific factors. Overall, our findings suggest that the mechanisms linking high automation risk to health are gender- and context-sensitive, and are shaped by broader economic conditions and workplace environments." (Author's abstract, IAB-Doku) ((en))

Zitatform

Vivarelli, Marco & Guillermo Arenas Diaz (2025): Innovation and employment in the era of artificial intelligence. In the face of AI revolution, concerns about possible technological unemployment should be aware of the complex and mixed employment impacts of technological change. (IZA world of labor 154,2), Bonn, o. S. DOI:10.15185/izawol.154.v2

Abstract

"The relationship between technology and employment has always been a source of concern, at least since the first industrial revolution. However, while process innovation can be job-destroying (provided that its direct labor-saving effect is not compensated through market mechanisms), product innovation can imply the emergence of new firms, new sectors, and thus new jobs (provided that its welfare effect dominates the crowding out of old products). Nowadays, the topic is even more relevant because the world economy is undergoing a new technological revolution centred on automation and the diffusion of Artificial Intelligence (AI)." (Author's abstract, IAB-Doku) ((en))

Zitatform

Vodă, Adina-Maria, Mihai Ciobotea, Doina Badea, Monica Roman & Marian Stan (2025): The Interactions Between Digitalization, Innovation and Employment in European Companies: Insights from a Latent Class Analysis. In: Economies, Jg. 13, H. 4. DOI:10.3390/economies13040104

Abstract

"There is increasing concern regarding the association between technological change and jobs. This study explores how different patterns of digitalization and innovation relate to job creation in European companies. We use data from the European Company Survey 2019 collected by Eurofound and Cedefop. We apply Latent Class Analysis (LCA) to identify the typologies of companies, mainly based on their level of technology adoption, innovation practices and employment patterns. We showcase four distinct classes of companies: moderate adoption of digital technology and strong international orientation, traditional and local, medium digitalization, process innovative with local focus and digital leaders and innovators, with specific patterns regarding digitalization, innovation and job creation. The digital leaders and innovators class revealed a high level of digitalization and innovation and maintained stable employment levels, with increased investments in staff training and tendency towards automation. Conversely, less-digitalized traditional companies are more susceptible to stagnation or employment decline. In general, the employment outlook is stable, without significant employment growth, signaling the need for balanced investments in innovation and digitalization that stimulate more and better jobs. This is the first study to apply LCA to explore complex relationships between digitalization, innovation, foreign trade, training investments and employment trends and offers fresh insights into company views towards employment in the digital era." (Author's abstract, IAB-Doku) ((en))

Zitatform

Wang, Ping & Tsz-Nga Wong (2025): Artificial Intelligence and Technological Unemployment. (NBER working paper / National Bureau of Economic Research 33867), Cambridge, Mass, 53 S.

Abstract

"How large is the impact of artificial intelligence (AI) on labor productivity and unemployment? This paper introduces a labor-search model of technological unemployment, conceptualizing the generative aspect of AI as a learning-by-using technology. AI capability improves through machine learning from workers and in turn enhances their labor productivity, but eventually displaces workers if wage renegotiation fails. Three distinct equilibria emerge: no AI, some AI with higher unemployment, or unbounded AI with sustained endogenous growth and little impact on employment. By calibrating to the U.S. data, our model predicts more than threefold improvements in productivity in some-AI steady state, alongside a long-run employment loss of 23%, with half this loss occurring over the initial five-year transition. Plausible change in parameter values could lead to global and local indeterminacy. The mechanism highlights the considerable uncertainty of AI's impacts in the presence of labor-market frictions. In the unbounded-AI equilibrium, technological unemployment would not occur. We further show that equilibria are inefficient despite adherence to the Hosios condition. By improving job-finding rate and labor productivity, the optimal subsidy to jobs facing the replacement risk of AI can generate a welfare gain from 26.6% in the short run to over 50% in the long run." (Author's abstract, IAB-Doku) ((en))

Zitatform

Warning, Anja (2025): Volkswirtschaftliche Perspektiven auf Künstliche Intelligenz. Rahmenbedingungen, Erwartungen, Befunde. In: Hans-Böckler-Stiftung (Hrsg.) (2025): Drei Denkwerkstätten - Neue Perspektiven für die Arbeit der Zukunft, Düsseldorf, S. 82-89.

Abstract

"KI und Generative KI bieten im aktuellen Strukturwandel hin zu einer stärker dienstleistungs- und wissensbasierten Ökonomie Möglichkeiten und Perspektiven wie keine technologische Entwicklung zuvor. Ob die erwarteten Produktivitätseffekte allerdings eintreten, hängt von verschiedenen Faktoren ab, auch von datenschutzrechtlichen Regelungen. Diese können Sicherheit bei der Nutzung von KI bedeuten, aber auch die Anwendungs- und Analysemöglichkeiten einschränken. Außerdem gilt es zu bedenken, dass KI-Anwendungen viele Aufgaben effizienter erledigen können, aber auch Energie und Ressourcen in Form von Strom, Servern, Rechenzentren und Leitungskapazität verbrauchen. So können sie die Green Economy unterstützen, belasten aber ihrerseits die Umwelt. Zudem können KI-Anwendungen zwar in Sekundenschnelle Analysen durchführen und Entscheidungen treffen, reproduzieren dabei aber vorhandenes Wissen und folgen den (Mainstream-)Denk- und Verhaltensmustern, mit denen sie trainiert wurden. Innovative und bahnbrechende Ideen werden wohl weiterhin der menschlichen Kreativität und dem menschlichen Erfindergeist vorbehalten bleiben." (Textauszug, IAB-Doku)

Zitatform

Weber, Enzo (2025): Wenn Ihr Job von KI betroffen ist, kann das eine große Chance sein. Gastbeitrag. In: Frankfurter Allgemeine Zeitung H. 12.02.2025 Frankfurt am Main.

Abstract

"Künstliche Intelligenz ersetzt immer mehr Arbeitsplätze. Sie dringt in mehr und mehr Tätigkeitsbereiche vor. Mittlerweile sind auch viele hoch qualifizierte Jobs betroffen – Ärzte, Rechtsanwältinnen, Journalisten und viele andere. Wir müssen uns diesem Wandel anpassen." (Autorenreferat, IAB-Doku, © Frankfurter Allgemeine Zeitung)

Zitatform

Weber, Enzo & Gerd Zika (2025): Wegen KI könnten 800.00 Arbeitsplätze wegfallen und neu entstehen. Gastbeitrag. In: Frankfurter Allgemeine Zeitung H. 19.11.2025 Frankfurt am Main.

Abstract

"Wie wird KI die Arbeitswelt verändern? Eine neue Studie wagt eine Prognose. Die zentrale Botschaft: Die Zahl der Arbeitsplätze bleibt weitgehend stabil, doch die Umwälzungen dahinter werden gewaltig sein." (Autorenreferat, IAB-Doku, © Frankfurter Allgemeine Zeitung)

Zitatform

Widuckel, Werner & Lutz Bellmann (2025): Nachhaltigkeit, berufliche Bildung und Personalentwicklung unter Berücksichtigung von KI. In: K. Gondlach, B. Brinkmann, M. Brinkmann & J. Plath (Eds.) (2025): Regenerative Zukünfte und künstliche Intelligenz, S. 73-82. DOI:10.1007/978-3-658-46577-3_7

Abstract

"Wie in vielen Ländern wird in Deutschland für den Klimaschutz umgebaut, um eine größere Energieeffizienz zu erreichen und mehr erneuerbare Energien einzusetzen. Es geht um nachhaltiges Wirtschaften, das nicht nur auf das Ziel der Gewinnmaximierung ausgerichtet ist, sondern auch soziale und ökologische Entwicklungsziele berücksichtigt. Betriebswirtschaftliche Vorteile wie Absatz- und Imagesteigerung, eine höhere Attraktivität als Arbeitgeber oder die Optimierung von Arbeitsabläufen sind mit höherem Gewinn und/oder niedrigeren Kosten verbunden. Künstliche Intelligenz (KI) kann z. B. in der Instandhaltung bei der Produktion, der Planung und der Qualitätssicherung mit wesentlichen Verbesserungen verbunden sein (Koch et al. 2022). Über das einzelne Unternehmen hinaus ist die gesamte Wirtschaft betroffen, die vor umwälzenden Herausforderungen wie der Dekarbonisierung von Dienstleistungs-, Produktions- und Reproduktionsprozessen, der Umstellung des Verkehrs und der Logistik oder dem Aufbau und Ausbau eines klimafreundlichen Gebäudemanagements steht (Messner 2022, S. 21)." (Autorenreferat, IAB-Doku, © Springer Nature)

Zitatform

Williamson, Harry, Dermot Coates, Kevin Daly, Keith FitzGerald & Neil Gannon (2025): Occupational exposures, complementarity and the potential consequences of A.I. for the labour market: some evidence from Ireland. In: Journal for labour market research, Jg. 59. DOI:10.1186/s12651-025-00418-w

Abstract

"The adoption of AI technology by industry could significantly disrupt our current understanding of “typical” economic activity. As AI comes to pervade more sectors and occupations over time, it is likely that this technology will give rise to challenges and risks but also opportunities and benefits. There is, however, a significant degree of uncertainty regarding how future waves of technological change will impact the economy, including the labour market. Recent research has found that 40% of employment globally is exposed to AI and that this rises to 60% of employment in advanced economies. We analyse exposure and complementarity in tandem in order to better understand the potential impact across occupation types in Ireland. We find that Ireland is relatively more exposed to AI than is the case for other advanced economies. We also find find that female workers in Ireland are more likely to work in highly exposed roles compared to males, that younger Irish workers are more exposed to AI than are older workers, and that both exposure complementarity to AI increase in line with educational attainment. Finally, we contend that the extent to which AI augments, or replaces, human labour in the medium to long-run will depend on a variety of economic, social and policy factors, including levels of AI regulation." (Author's abstract, IAB-Doku) ((en))

Zitatform

Winters, Jutta & Jonathan P. Latner; David Autor (interviewte Person) (2025): Does automation replace experts or augment expertise? The answer is yes (Interview). In: IAB-Forum H. 09.01.2025. DOI:10.48720/IAB.FOO.20250109.01

Abstract

"David Autor, Professor of economics at the Massachusetts Institute of Technology (MIT), gives a Special Lecture at the IAB on 15 January 2025. In this accompanying interview, he discusses the impact of Artificial Intelligence on wages and employment, outlines the crucial role of expertise and gives insights on policy-approaches for supporting workers in rapidly changing labor markets." (Author's abstract, IAB-Doku) ((en))

Zitatform

Wohlrabe, Klaus (2025): Mehr KI, weniger Jobs? Was Unternehmen in Deutschland erwarten. In: ifo Schnelldienst digital, Jg. 6, H. 8, S. 1-12.

Abstract

"Die Nutzung von Künstlicher Intelligenz (KI) in der deutschen Wirtschaft nimmt weiter zu – insbesondere in größeren Unternehmen. Auf Basis von Sonderfragen im Rahmen der monatlichen Konjunkturumfragen wird gezeigt: Derzeit fallen die direkten Beschäftigungseffekte von KI noch gering aus. Für die kommenden fünf Jahre wird jedoch deutlich häufiger ein Einfluss auf die Beschäftigtenzahl erwartet – vor allem in Form eines Stellenabbaus. Die Einschätzungen unterscheiden sich je nach Branche teils erheblich. Trotz zunehmender Praxiserfahrung bleibt eine deutliche Unsicherheit über die langfristigen Arbeitsmarktauswirkungen bestehen. Die Ergebnisse liefern einen ersten Einblick in die beschäftigungspolitischen Erwartungen von Unternehmen in Deutschland mit Blick auf KI." (Autorenreferat, IAB-Doku)

Zitatform

Wood, Alex J. (2025): Contesting Algorithmic Workplace Regimes in an Era of Flexible Despotism. In: Sociology, S. 1-18. DOI:10.1177/00380385251388255

Abstract

"This article investigates recent claims that the growing use of algorithms is giving rise to a novel workplace regime. The article makes two conceptual contributions: first, it identifies the generic characteristics of this supposed algorithmic workplace regime. Second, it puts into question the exceptionalism of algorithmic workplace regimes. This is achieved by bringing the centrality of non-algorithmic management techniques in co-constituting algorithmic regimes into focus and by historically situating the regime’s emergence within the wider workplace regime literature. In doing so the article questions the novelty and distinctiveness of algorithmic workplace regimes, arguing that such regimes are better understood as a subtype of flexible despotism that can be traced back to the 1980s." (Author's abstract, IAB-Doku) ((en))

Zitatform

Yuan, Bocong, Jiannan Li & Hairong Zhao (2025): Education pathways to mitigate automation anxiety: skill development as key for job satisfaction in the age of machines replacing human. In: International Journal of Manpower, Jg. 46, H. 9, S. 1676-1698. DOI:10.1108/ijm-02-2024-0093

Abstract

"Purpose: The application of intelligent machine in the workplace has led to increasing concern about technically induced unemployment. This study is to investigate the mechanism of how such risk affects the job satisfaction. Design/methodology/approach: We use the secondary data from SHARE (wave 8) and a longitudinal survey to examine the influence mechanism of how intelligent machine job substitution risk affects job satisfaction. Findings: Results show that intelligent machine job substitution risk has a negative impact on job satisfaction. Besides, skill development opportunity mediates the negative relation between intelligent machine job substitution risk and job satisfaction. Further, work support buffers the negative relation between intelligent machine job substitution risk and skill development opportunity, while enhancing the positive relation between skill development opportunity and job satisfaction. Originality/value: This study is the first to examine the mediation role of skill development opportunity in the relation between the intelligent machine job substitution risk and job satisfaction. Also, this study is the first to explore the role of work support in the above relation. This study enriches relevant research regarding the intelligent machine application in workplace and provides important insights for organization management." (Author's abstract, IAB-Doku, © Emerald Group) ((en))

Zitatform

Zeng, Shihong, Zeyuan Yang & Zhibin Zhang (2025): Servitization and employment effects in manufacturing: international experiences under digital technology conditions. In: Applied Economics, S. 1-18. DOI:10.1080/00036846.2025.2526854

Abstract

"This study examines the impact of manufacturing servitization on employment in 43 countries between 2000 and 2020, considering the role of digital technology. Using data from the World Input-Output Tables and the International Labour Organization, we explore how servitization affects employment in manufacturing and productive services. Our findings show that servitization leads to a substitution effect on manufacturing employment and a creation effect on productive service employment. The adoption of digital technology amplifies the substitution effect while weakening the creation effect. We further conduct heterogeneity analysis based on factor demand, factor input, level of industrialization, and level of economic development. We find that the substitution effect is most significant in low-tech manufacturing, while the creation effect is strongest in high-tech manufacturing. Regarding factor input, the creation effect is concentrated in low- to mid-tech productive service sectors. The substitution effect is more pronounced in less industrialized and developed countries, while the creation effect is more evident in highly industrialized and developing countries.The study identifies two key mechanisms through which servitization influences employment, namely improving information technology levels and reducing institutional distance. These findings contribute to the understanding of employment transitions in the digital economy and provide valuable insights for policymaking." (Author's abstract, IAB-Doku) ((en))

Zitatform

Zeyer-Gliozzo, Birgit (2025): Returns to formal, non-formal, and informal further training for workers at risk of automation. In: Journal of education and work, Jg. 17. DOI:10.1080/13639080.2024.2447037

Abstract

"The automation of job tasks due to technological change increases the pressure on workers whose jobs consist largely of such activities. In this context, politics and science attach great importance to further training, although the benefits for affected workers have hardly been investigated. Drawing on human capital theory and the task-based approach, this study examines the effect of further training on job tasks and the probability of automation for workers at risk of automation. These are workers who (a) work in routine intense jobs or (b) have a high probability of automation, based on the expert assessment of the automatability of occupations by Frey and Osborne (2017). Using data from the German National Educational Panel Study (NEPS), fixed-effects models are estimated to account for unobserved heterogeneity. The results show that informal further training is most effective in increasing non-routine tasks and reducing routine tasks and the probability of automation. Formal and non-formal training also show returns in some cases, although these are often not significant. The results thus suggest that not all forms of further training are helpful in adapting to new demands in the course of technological change and provide starting points for future research." (Author's abstract, IAB-Doku) ((en))

Zitatform

Zhang, Su, Xiaolin Wang, Yan Xia & Huijuan Wang (2025): The AI Redemption: How technology is rewriting the rules of cross-industry labor mobility. In: International Review of Economics and Finance, Jg. 103. DOI:10.1016/j.iref.2025.104575

Abstract

"This study considers the evolution and iteration of digital technology, conducting both theoretical and empirical research on the effects of information technology and artificial intelligence technology on cross-industry labor mobility. Theoretically, we construct a general equilibrium model that includes labor and digital technology to analyze the intrinsic mechanisms by which digital technology affects cross-industry labor mobility. Empirically, using the probit model and the instrumental variable approach, we find robust evidence of a significant positive effect of digital technology on cross-industry labor mobility through the pooled four-wave data from the China Family Panel Studies (CFPS) from 2014 to 2020. The findings indicate that digital technology significantly promotes cross-industry labor mobility. Mechanism analysis reveals that information technology, represented by computers, drives low-skilled labor towards non-skill-intensive industries through substitution and productivity effects, while artificial intelligence technology promotes the flow of both low-skilled and high-skilled labor towards skill-intensive industries through “de-skilling” and “re-skilling”. The impact of digital technology on cross-industry labor mobility varies significantly across different genders, the type of hukou, age, and employment types. Further mechanism analysis suggests that digital technology facilitates higher wage gains by promoting cross-industry labor mobility." (Author's abstract, IAB-Doku, © 2025 The Authors. Published by Elsevier Inc.) ((en))

Zitatform

Zhou, Yongguang, Weihong Xie, Jingwu Li & Qun Li (2025): Effects of digital innovation on income inequality among different workforces: evidence from Chinese industries. In: Applied Economics, Jg. 57, H. 22, S. 2809-2821. DOI:10.1080/00036846.2024.2331424

Abstract

"To understand the impact of digital innovation on the workforce and its role in achieving common prosperity, this paper uses data from Chinese A-share listed companies during 2006–2021 to investigate the effects of digital innovation on income inequality among different industry-level groups. We find that digital innovation significantly reduces income inequality among employees across industries, but it does not significantly impact income inequality within management groups. Through mechanistic analysis, we find that digital innovation decreases income inequality among ordinary employees whose incomes are closely linked to company performance and thereby for the entire workforce by narrowing the income gap across industries. However, as digital innovation does not significantly influence evaluation systems (e.g. educational degrees) for management income, it does not contribute to reducing income inequality among managerial levels. These findings provide valuable insights to develop policies for common prosperity." (Author's abstract, IAB-Doku) ((en))

Zitatform

Zika, Gerd, Theresa-Marie Hassemer, Markus Hummel, Bennet Krebs, Tobias Maier, Anke Mönnig, Christian Schneemann, Enzo Weber & Johanna Zenk (2025): Künstliche Intelligenz: Potenzielle Effekte für den deutschen Arbeitsmarkt. (IAB-Forschungsbericht 23/2025), Nürnberg, 58 S. DOI:10.48720/IAB.FB.2523

Abstract

"Die fortschreitende Entwicklung und Integration von Künstlicher Intelligenz (KI) prägt zunehmend die globalen Märkte und Arbeitsweisen und hat auch in Deutschland einen immer größeren Einfluss. Deshalb wird mit der vorliegenden Szenarioanalyse für einen Zeitraum von 15 Jahren untersucht, welchen Einfluss KI auf die Entwicklung von Wirtschaft und Arbeitsmarkt in Deutschland entfalten könnte. Um Aussagen über potenzielle Effekte eines KI-Szenarios treffen zu können, müssen entsprechende Annahmen getroffen werden. Anschließend wird das KI-Szenario mit einem Referenz-Szenario verglichen. Im KI-Szenario wird unterstellt, dass die notwendige Infrastruktur für die KI-Entwicklung und -Integration vorhanden ausgebaut wird. Insbesondere muss ausreichend Rechenkapazität in Rechenzentren verfügbar sein. Zudem wird KI je nach Branche unterschiedlich stark zum Einsatz kommen. So müssen von den Betrieben der jeweiligen Branchen unterschiedlich hohe Vorlaufinvestitionen getätigt werden. Dabei ist zu erwarten, dass die Wirkung des KI-Einsatzes je nach Branche unterschiedlich ausfallen wird. In manchen Branchen können Materialeinsparungen, in anderen Produktivitätssteigerungen oder neue Absatzmöglichkeiten realisiert werden. In vielen Branchen dürften diese Effekte auch in Kombination auftreten. Schließlich entwickeln auch andere Länder KI und wenden diese an. Für das KI-Szenario wird deshalb angenommen, dass die Wirkungen im Ausland insgesamt ähnlich ausfallen werden wie in Deutschland, mit den entsprechenden Effekten auf deutsche Importpreise. Den Berechnungen zufolge lässt sich durch KI insbesondere mittel- und langfristig eine deutlich höhere Wertschöpfung bei nahezu gleichem Arbeitseinsatz erzielen als im Referenz-Szenario. Unter den getroffenen Annahmen kann das jährliche Wirtschaftswachstum um durchschnittlich 0,8 Prozentpunkte höher liegen als im Referenz-Szenario. Kumuliert über die nächsten 15 Jahre könnten so 4,5 Billionen Euro an zusätzlicher Wertschöpfung erwirtschaftet werden. Die Zahl an Arbeitsplätzen liegt im KI-Szenario nach 15 Jahren insgesamt auf einem ähnlichen Niveau wie im Referenz-Szenario. Dennoch zeigen sich deutliche Verschiebungen am Arbeitsmarkt. So gibt es Wirtschaftsbereiche, in denen der Arbeitskräftebedarf steigt, während in anderen Bereichen Arbeitsplätze verloren gehen. Zu Beginn der verstärkten KI-Entwicklung und -Integration ist zunächst mit einem insgesamt höheren Arbeitskräftebedarf zu rechnen, um beispielsweise die benötigte Infrastruktur bereitzustellen, Daten zu erschließen oder Modelle zu entwickeln. Mittelfristig kommt es zu einem etwas geringeren Arbeitskräftebedarf aufgrund zunehmender Effizienzgewinne. Gegen Ende des Projektionszeitraums besteht jedoch Potenzial, diese Arbeitsplatzverluste wieder auszugleichen. Dabei zieht insbesondere die Entwicklung neuer Geschäftsmodelle einen höheren Arbeitskräftebedarf nach sich. Insgesamt gleichen sich die gegenläufigen Effekte auf die Beschäftigung weitgehend aus. Im KI-Szenario führt der Einsatz von KI unterm Strich also nicht ausschließlich zu positiven oder negativen Effekten auf dem Arbeitsmarkt. Vielmehr kommt es in einigen Bereichen zu Verschiebungen. Bei den IT- und Informationsdienstleistern liegt die Zahl der Erwerbstätigen nach 15 Jahren beispielsweise um rund 110.000 Personen höher als im Referenz-Szenario. Hingegen liegt die Zahl der Erwerbstätigen bei den Unternehmensdienstleistern im KI-Szenario zum gleichen Zeitpunkt um rund 120.000 Personen niedriger. Eine KI-induzierte Absenkung der Erwerbstätigenzahlen ist jedoch nicht zwingend mit einer Verschlechterung der Arbeitsmarktlage verbunden. Vielmehr könnten knappe Personalressourcen langfristig effizienter eingesetzt werden, wodurch Potenzial besteht, etwaige Arbeitskräfteengpässe in anderen Bereichen zu reduzieren. Gleichzeitig ist zu erwarten, dass ein verstärkter KI-Einsatz die Anforderungen an die Beschäftigten verändern wird. So dürften insbesondere hochqualifizierte Tätigkeiten stärker von KI-Anwendungen betroffen sein als von Entwicklungen im Bereich klassischer nicht-lernender Software." (Autorenreferat, IAB-Doku)

Zitatform

Zou, Wei, Liming Zheng & Ruiqi Ma (2025): Can inequality and intergenerational upward mobility coexist: the impact of skill-biased technological change. In: Applied Economics Letters, S. 1-8. DOI:10.1080/13504851.2025.2550571

Abstract

"We extend the Maoz and Moav (1999) model by introducing a CES production function to examine the impact of skill-biased technological change (SBTC) on inequality and intergenerational upward mobility. Our study shows that (1) when the production function shifts from a Cobb-Douglas form to a CES form, high-skilled workers rapidly converge to a unique steady state; (2) while SBTC exacerbates inequality, it raises the steady-state number of high-skilled workers, establishing a positive correlation between inequality and intergenerational upward mobility." (Author's abstract, IAB-Doku) ((en))

Zitatform

ten Berge, Jannes & Fabian Dekker (2025): New technology and workers’ perceived impact on job quality: Does labor organization matter? In: Economic and Industrial Democracy, Jg. 46, H. 2, S. 619-654. DOI:10.1177/0143831x241265911

Abstract

"There is an emerging literature focusing on the impact of technological change on work quality. This study contributes to the literature by examining (1) workers’ expectations regarding the effect of technological change on perceived job insecurity, as well as physical and psychological job demands, and (2) how these expectations are shaped by the degree of labor organization within countries. The article uses cross-national data for 25 OECD countries. It is found that labor organization decreases perceived levels of job insecurity related to technological change, but also lowers workers’ expectations of technology improving the quality of their work. These findings may indicate that in environments where technological change is less strongly moderated by organized labor, workers put greater emphasis on technology as a driver of (short-term) work changes. Alternatively, these findings may signal a lack of ‘worker power’ of organized labor to enforce technologies that improve the quality of employment." (Author's abstract, IAB-Doku) ((en))

Zitatform

(2025): Erprobung neuer Methoden der Mustererkennung bei der Statistik der BA. Vorhaben im Rahmen des BMAS-Datenlabors. (Grundlagen: Methodenbericht / Bundesagentur für Arbeit), Nürnberg, 81 S.

Abstract

"Die Statistik der BA betrachtet Methoden der Künstlichen Intelligenz, insbesondere des Maschinellen Lernens, als sinnvolle Ergänzung ihres Methodenspektrums zur Weiterentwicklung in konkreten Aufgabenbereichen. Einsatzfelder bestehen vor allem dort, wo lernfähige Algorithmen die automatische Verarbeitung großer Datenmengen mit dem Ziel der Mustererkennung ermöglichen. Vielversprechend erscheint der Einsatz dieser Methoden unter anderem in zwei Bereichen: Bei der Unterstützung der bestehenden Validierung von Daten im Zuge der Stichtagsverarbeitung sowie bei der Zuordnung von Beschäftigungsbetrieben zu Wirtschaftszweigen. Machbarkeitsanalysen für beide Anwendungen wurden durch das BMAS-Datenlabor gefördert und aus Mitteln der Europäischen Union „NextGenerationEU“, die im Rahmen des Deutschen Aufbau- und Resilienzplans (DARP) bereitgestellt wurden, finanziert. Das entsprechende Vorhaben der Statistik der BA, das „Datenlabor Statistik“, hatte eine Laufzeit von November 2022 bis Dezember 2024. Die beiden genannten Einsatzfelder bildeten zwei Module eines Gesamtvorhabens. Im Modul „Validierung“ wurden neue Methoden zur Validierung großer Mengen statistischer Daten konzipiert, exemplarisch eingesetzt und bewertet, mit deren Hilfe sich im Vergleich zum aktuellen Verfahren erheblich mehr Daten hinsichtlich auffälliger Konstellationen prüfen lassen. Die als erfolgversprechend bewerteten neuen Ansätze adressieren bisher weniger beachtete Bereiche der Plausibilitätsprüfung: fachstatistikübergreifende Analysen, die Berücksichtigung von Merkmalskombinationen sowie Analysen auf tieferer regionaler Ebene. Ein Ansatz für die beiden erstgenannten Bereiche wurde unmittelbar nach dem Datenlabor in einen testweisen Produktivbetrieb und in weitere Analysen überführt. Die geprüften Ansätze zur Erkennung fehlerhafter schleichender Veränderungen sowie für Einzelfallprüfungen auf tieferen Schichten der Verarbeitung haben sich als weniger aussichtsreich herausgestellt; hier sind weitere Analysen nötig. Im Modul „Wirtschaftszweige“ wurde vor allem ein Lösungsansatz als erfolgversprechend und praktisch umsetzbar bewertet: Die Zuordnung von Beschäftigungsbetrieben zu Wirtschaftszweigen auf Basis eines speziell trainierten großen Sprachmodells, das Freitext-Beschreibungen zu Betriebszwecken verarbeitet. Dieser Ansatz soll nun in der Praxis umgesetzt werden. Wertvolle Erkenntnisse wurden in diesem Modul zudem für Lösungsansätze auf Basis der folgenden Datenquellen gewonnen: Stellenanzeigen, Betriebsbezeichnungen und Informationen zu den Tätigkeiten von Beschäftigten. Zu den übergreifenden Ergebnissen des Datenlabors der Statistik zählen der Wissensgewinn durch methodischen Austausch innerhalb und außerhalb der BA, das Erschließen bisher nicht genutzter technischer Architekturen sowie leistungsfähiger Hardware und Entwicklungswerkzeuge." (Autorenreferat, IAB-Doku)

Zitatform

(2025): How do people experience new technologies and generative AI? Insights from a few countries worldwide. (OECD Policy Insights on Well-being, Inclusion, Sustainability and Equal Opportunity 23), Paris, 12 S. DOI:10.1787/49b8d10e-en

Zitatform

(2025): D21-Digital-Index 2024/2025. Jährliches Lagebild zur Digitalen Gesellschaft. (D21-Digital-Index), Berlin, 41 S.

Abstract

"Der D21-Digital-Index ist Deutschlands wichtigstes Lagebild zur Digitalisierung der Gesellschaft. Er zeigt, wie tief die digitale Transformation verschiedene Lebensbereiche durchdringt und wie gut Bürger*innen mit den Anforderungen des Wandels umgehen können. Gleichzeitig offenbart er Spaltungen und Herausforderungen: Wer profitiert, wer droht abgehängt zu werden? Der D21-Digital-Index ist mehr als eine Analyse: Er ist Basis für wirkungsvolles Handeln von Politik, Wirtschaft und Gesellschaft, um die Resilienz im digitalen Wandel zu stärken und gemeinsam eine inklusive digitale Zukunft zu gestalten, in der alle von den Chancen der Digitalisierung profitieren. Im Fokus steht in diesem Jahr der digitale Wandel als Treiber für weitere Transformationsprozesse, etwa bei der Wertschöpfung, beim Informations- und Kommunikationsverhalten oder beim ökologischen Wandel. Denn unsere Gesellschaft befindet sich inmitten eines tiefgreifenden Wandels. Der D21-Digital-Index begleitet sie in dieser Transformation seit vielen Jahren. Ein weiterer Schwerpunkt der Studie liegt auf der digitalen Bildung und den Kompetenzen, die für eine souveräne und kritische Nutzung digitaler Technologien erforderlich sind. Zudem wird untersucht, wie die Bevölkerung den Einsatz Künstlicher Intelligenz wahrnimmt: Welche Erwartungen, Chancen und Bedenken gibt es? Welche Faktoren beeinflussen die Akzeptanz neuer KI-Technologien im Alltag? Wie steht es um das Bewusstsein für die Transformationskraft von KI, insbesondere für die Beschäftigungschancen?" (Autorenreferat, IAB-Doku)

Zitatform

(2025): OECD Skills Outlook 2025. Building the Skills of the 21st Century for All. (OECD skills outlook), Paris, 261 S. DOI:10.1787/26163cd3-en

Abstract

"The OECD Skills Outlook 2025 examines how countries can build the 21st-century skills needed to sustain growth and social progress. It explores how differences in background, education and opportunity shape who develops, uses and benefits from key skills such as literacy, numeracy and adaptive problem solving. Differential access to skills development limits the potential of many, thereby constraining economic performance. Socio-economic background strongly influences who builds skills that are valued in the labour market, whereas differences between men and women appear mainly in how skills are used and rewarded. As skill demands evolve faster than policy cycles, investing in lifelong learning and using timely labour-market intelligence are crucial to help people adapt, strengthen productivity and ensure that no one is left behind in a changing world." (Author's abstract, IAB-Doku) ((en))