The workshop, titled “The AI Privacy Paradox: Balancing Data Minimization with Large Language Models,” explores the “legal-technological” paradox arising from the GDPR’s strict requirement for data minimization (which mandates that personal data be used only to the extent strictly necessary) and the AI Act’s comparatively permissive stance. Participants will analyze the practical implications of this divergence, specifically how companies increasingly invoke a “legitimate interest” in technological development to justify opt-out policies that may elude GDPR consent standards. Through a deep dive into the NOYB vs. Meta case, students will examine the tension between massive data harvesting machine learning and individual rights such as access and erasure. The session culminates in an interactive evaluation of the European Commission’s Digital Omnibus and the proposed Article 88c.