MDI Impact Brief — Resolving the Tension between Privacy and Generative AI
Generative AI tools and applications are advancing at an astonishing rate. From generating code, email, and music to supporting decision making, generative AI is becoming common in everyday workflows. Notwithstanding the apparent strengths, the weaknesses are more subtle, including biases against diverse subpopulations, deceptive content generation, and the use of private (or at least not public) data during model construction. It is the latter that we are going to discuss in this impact brief.
This Impact Brief explores the question — how do we resolve the tension between improving the quality of generative AI tools, including publicly available generative models, and protecting our privacy? Outlining suggestions of how to address generative AI related privacy risks, as well as describing some initiatives and regulations that are underway, the brief outlines five principles that policymakers should keep in mind as they consider different policies and regulations for generative AI that can help reduce privacy harms in LLMs. Some of these principles can be used to define generative AI policy and promote a new generation of AI tools that consider privacy when incorporating more sensitive data into their tools.
MDI Impact Brief Contributors
Eleni Antoniadou Dimokidis, Ph.D., MDI Non-Resident Affiliate and Head of Healthcare Technology across Asia Pacific and Japan at Amazon Web Services
Renée DiResta, Associate Research Professor with the Massive Data Institute and Tech & Public Policy at the McCourt School of Public Policy at Georgetown University
Andrew Gamino-Cheong, MDI Non-Resident Affiliate and Co-founder & CTO of Trustible
Lisa Singh, Ph.D., MDI Director, Sonneborn Chair, Chair and Professor in the Department of Computer Science, and Professor at the McCourt School of Public Policy at Georgetown University
About the Massive Data Institute
The Massive Data Institute (MDI) at the McCourt School of Public Policy at Georgetown University is an interdisciplinary research institute that connects experts across computer science, data science, public health, social science and public policy to tackle societal scale issues and impact public policy in ways that improves people’s lives through responsible evidence-based research.