Social media dependency with the precision of an algorithm: How the EU AI Act fails to tackle profiling and micro-targeted content

Resumo

This paper critically examines the EU Artificial Intelligence Act (AI Act) in the context of its ability to regulate AI-driven profiling and micro-targeting. It explores the implications of how social media algorithms exploit user data, cognitive biases and emotional vulnerabilities to influence behaviour and reinforce inequalities. In an analytical part, this paper highlights the limitations of the AI Act’s Article 5 in addressing manipulative AI practices and its limited scope covering vulnerable groups. It argues that the provision adopts a narrow and static conception of vulnerability, failing to capture broader behavioural and social forms of dependency. Finally, the paper argues that these deficiencies hinder the AI Act’s effectiveness in limiting algorithmic manipulation and proposes more robust obligations on transparency, meaningful disclosure, and user-centred explainability to reinforce user autonomy and mitigate digital dependency.