European Workplaces Turn to AI for Employee Decisions as Worker Rights Lag Behind
Amsterdam, Friday, 20 February 2026.
One in four European companies now use artificial intelligence to make critical decisions about their employees’ careers, from hiring and scheduling to performance evaluations and surveillance. This dramatic shift has caught labor unions off-guard, with many struggling to negotiate worker protections against algorithmic management systems that can monitor mental health data and track employees even when offline. While the EU passed some platform worker protections in 2024, unions are pushing for comprehensive AI workplace legislation by 2026 as sophisticated algorithms become increasingly pervasive across European organizations.
The Scope of Algorithmic Management in European Workplaces
A comprehensive study conducted by Visionary Analytics for the European Commission revealed that algorithmic decision-making systems are now deployed across multiple facets of employee management [1]. These AI-powered systems are being used for recruitment processes, task scheduling, worker direction, surveillance activities, and performance evaluations [1]. The technology extends far beyond simple automation, with algorithms now determining fundamental working conditions including hours, wages, shift assignments, and performance assessments [1]. Perhaps most concerning for privacy advocates, these systems can collect mental health data and continue tracking employees even when they have logged off from work systems [1].
Union Challenges in the Digital Age
Labor organizations across Europe are encountering significant obstacles in addressing this technological transformation. Tea Jarc, confederal secretary of the European Trade Union Confederation (ETUC), explained the fundamental challenge: “It’s a new topic, it’s a challenging topic, and many of the trade unions do not actually have the expertise or the capacities to address it” [1]. The ETUC reports particular difficulty in negotiating collective agreements that adequately address algorithmic management in workplaces [1]. This expertise gap has led to enforcement of collective agreements being “very low” according to union reports [1]. Despite these challenges, some progress has emerged, with Uni Europa reporting that 23 agreements mentioning algorithmic management have been signed throughout Europe [1].
Legal Framework and Corporate Responsibilities
The legal landscape surrounding AI workplace implementation varies significantly across European jurisdictions. In the Netherlands, Paul Snijders of WS Advocaten & Mediators explains that employers generally have the authority to mandate AI tool usage under Article 7:660 of the Burgerlijk Wetboek, which requires employees to follow employer instructions related to work performance and company order [2]. However, this authority comes with important limitations - any AI implementation involving personal data processing falls under GDPR regulations, requiring legal basis and balancing business interests against employee privacy [2]. Snijders emphasizes that “not every smart tool that can technically do something is also legally allowed” [2]. When AI applications affect working conditions, privacy, or monitoring, Dutch law requires approval from the ondernemingsraad (works council) under Article 27 of the Works Council Act [2].
The Growing Sophistication and Future Trajectory
Industry experts anticipate a significant expansion in AI workplace applications over the coming decade. Alessio Bertolini, a researcher at the Oxford Internet Institute in the United Kingdom, predicts that “what we’re going to see more in the next few years is only a more sophisticated version [of algorithms] that [are] definitely more pervasive in the workplace” [1]. The study indicates that the use of algorithms and AI in workplaces is expected to “grow rapidly in the next 10 years” [1]. However, recent research from Carnegie Mellon University suggests limitations in current AI capabilities, with even the best-performing AI model, Claude 3.5 Sonnet, completing only 24 percent of workplace tasks fully when tested in a simulated company environment [8]. This research indicates that while AI can serve as a valuable assistant, a fully AI-managed enterprise remains unfeasible with current technology [8]. Tea Jarc of the ETUC emphasizes the need for corporate transparency, stating: “It’s time that those companies understand that they cannot hide [management decisions] behind the algorithms” [1].