AI-PHI - Session 22 - Beyond Consent, The Limits of Data Agency in Generative AI Systems - Giada Pistilli

​# Beyond Consent: The Limits of Data Agency in Generative AI Systems

​Can traditional notions of consent meaningfully apply to AI-generated content? What happens when our data creates representations of ourselves beyond our control? I will examine how generative AI systems challenge conventional legal and ethical frameworks built around consent, analyzing how our understanding of consent becomes problematic when applied to AI systems that generate content from personal data. What does it mean to consent in an environment where outcomes cannot be foreseen? I will explore three fundamental challenges: the scope problem, temporality problem, and autonomy trap, which collectively create a "consent gap" in AI systems. Through concrete examples, I will illustrate how current legal frameworks inadequately address these challenges. So, how might we reconceptualize consent for the age of generative AI? Should responsibility shift from individuals to system developers? What alternative ethical frameworks might better protect our rights and autonomy? The talk will conclude by examining these questions that emerge when we move beyond traditional notions of consent in our increasingly AI-driven world.

​## About Giada Pistilli

​Giada Pistilli is a philosopher specialized in AI ethics, focusing on conversational agents and generative AI. With a PhD in philosophy from Sorbonne University, her research explores the intersections of value systems, ethical frameworks, multilingualism, and cultural biases in emerging technological landscapes. As Principal Ethicist at Hugging Face, she conducts interdisciplinary research that bridges philosophical inquiry with technological development. In this role, she also oversees the platform's content moderation policy, applying ethical principles to the challenges of large-scale AI systems.