AI, Privacy Law and Machine Unlearning

In this episode of the Discovery podcast, Jevan Hutson, J.D. '20, incoming assistant professor of law and director of the Technology Law & Public Policy Clinic, discusses the article "Forget Me Not? Machine Unlearning's Implications for Privacy Law," which he co-authored in The Columbia Science & Technology Law Review. The article explores the intersection of generative AI, privacy and data protection law and focuses on "machine unlearning," an emerging tool designed to address privacy concerns. As generative AI systems increasingly process vast amounts of personal and sensitive data, the challenges they pose to privacy regulations are intensifying.

Machine unlearning allows for the selective removal or suppression of specific data — such as personal information that individuals request be deleted — from AI models. This process aims to help organizations comply with legal obligations and policy goals, particularly in response to data subject requests for data deletion. Hutson assesses whether the legal and remedial aims of privacy laws can be reconciled with the technical capabilities of machine unlearning.

Listeners are invited to consider the future of privacy law in an AI-driven world and the role of machine unlearning in preserving privacy rights.