Rethinking Privacy in the AI Era: Policy Provocations for a Data-Centric World

Abstract

This paper examines how current and future privacy regulations impact AI development. It highlights the escalating demand for training data, posing privacy risks at both individual and societal levels. Existing regulations, while grounded in Fair Information Practices, fall short in regulating AI's data needs. To address these challenges, the paper suggests shifting to opt-in data collection, enhancing transparency in the AI data supply chain, and advocating for new governance mechanisms to empower individuals in managing their data.