HARPA - Human-centered Approaches to Responsible Privacy for and with AI

HARPA – Human-centered Approaches to Responsible Privacy for and with AI addresses the urgent need to protect privacy in an era dominated by Artificial Intelligence (AI). While AI offers vast potential to enhance productivity and innovation, it also magnifies privacy risks by processing personal data at scale, often without transparency or consent. HARPA views privacy not merely as a technical issue but as a cornerstone of autonomy, dignity, and human rights. Its mission is twofold: to mitigate privacy risks arising from AI and to harness AI itself to develop usable, human-centered privacy solutions. HARPA identifies four core challenges: fragmented knowledge, cross-border regulatory complexity, lack of capacity and inclusivity, and gaps in education and awareness. To tackle these, it will create a transnational, interdisciplinary network uniting experts in computer science, law, psychology, and humancomputer interaction. The Action’s objectives include building a conceptual framework for human-centered privacy, harmonizing methodologies, and analyzing recurring AI-privacy failures, as well as creating an open repository, training early-career researchers, supporting cross-country inclusiveness, and fostering interdisciplinary collaboration. The Action advances the state of the art by establishing a European knowledge infrastructure, harmonized frameworks for Privacy-by-Design, and practical educational resources. Networking is central to HARPA’s rationale, enabling sustained, cross-sectoral dialogue and inclusion of underrepresented stakeholders. With a critical mass of 30 proposers (27 from 18 full COST members, and 3 international proposers), HARPA will build capacity, raise awareness, and transform ethical principles into actionable, human-centric solutions — ensuring that privacy remains usable, equitable, and resilient in the age of AI.