When Machines Manage: Workers' Demands and Trade Union Responses to AI

The diffusion of artificial intelligence (AI) and related digital technologies is transforming the organization of work, yet their implications for labor-market institutions remain only partially understood. Existing research has largely focused on automation technologies, leaving open how complex forms of digitalization (e.g., algorithmic management and workplace surveillance) shape worker organization and collective bargaining. This project starts from the premise that technology matters not only because it can replace labor, but because it reorganizes how work is performed, monitored, and coordinated.

The project adopts a two-level empirical design. At the individual level, we use data from the European Social Survey (ESS), covering 15 countries between 2012 and 2024, to examine how exposure to different types of digital technologies is associated with union membership, working conditions, and political attitudes. We distinguish among displacement, augmentation, and monitoring technologies, and link these to occupational structures. The results point to heterogeneous effects: some technologies are associated with higher unionization and improved outcomes, while others correlate with weaker union attachment and more adverse working conditions. These findings suggest that digitalization reshapes both workers' demand for collective protection and their capacity to organize in ways that depend on how technologies alter control, autonomy, and job stability.

At the institutional level, the project builds and analyzes a novel, large-scale database of collective bargaining agreements (CBAs), covering Italy, Spain, France, Portugal, Finland, Canada, and the United States over the last decades. This cross-national corpus enables a systematic analysis of how technological change is reflected in contractual provisions across sectors, countries, and over time. We conceptualize CBAs as a central arena where the consequences of technological change are increasingly negotiated and codified, allowing us to trace how unions and employers respond to emerging forms of digitalization, including surveillance, data collection, and algorithmic management.

Methodologically, the project adopts a computational social science perspective, leveraging text-as-data approaches to analyze a large corpus of institutional documents. This strategy allows us to identify patterns in how technological change is addressed across agreements, sectors, and countries, moving beyond case-based evidence toward a more generalizable understanding.

Substantively, the project advances the political economy of automation by linking micro-level evidence on worker attitudes and behavior to institutional responses observed in collective bargaining. By integrating individual-level analysis with cross-national evidence on CBAs, it sheds light on how AI and digital technologies generate heterogeneous pressures on labor markets and how these pressures are mediated through labor institutions.