The increasing integration of artificial intelligence (AI) systems in the workplace has transformed labour processes by automating managerial functions such as task assignment, supervision, and worker evaluation. These developments exacerbate existing power asymmetries and create new vulnerabilities for workers.

In response to these challenges, among others, the European Union has developed the first comprehensive regulatory framework for AI; amongst its initiatives, the AI Act and the Platform Work Directive (PWD), both ratified in 2024, occupy a prominent position. However, the AI Act grants the AI industry self-assessment rights, limits external oversight and worker protection while failing to empower workers and their representatives to act as regulatory agents. Conversely, the PWD offers a broader understanding of algorithmic risks and includes provisions for proper worker classification and algorithmic transparency. However, its effectiveness is undermined by its non-binding nature and dependence on member states for implementation. Moreover, by focusing its stronger protections on platform workers, the EU risks further fragmenting labour relations concerning AI.

This paper critically examines the EU's regulatory framework, arguing that mainstream approaches reduce worker participation to symbolic gestures, assuming a convergence of interests between labour and capital. As an alternative, a workerist perspective is proposed, grounded in the presumption of conflict and treating technology as a contested domain. This approach seeks to empower workers and trade unions through a broader understanding of transparency, rights of information, veto powers, and collective agreements.