Governing the Unforesseable

As Artificial Intelligence (AI) becomes increasingly important in our everyday lives, governance mechanisms for AI are receiving increasing attention. At the same time, AI presents significant challenges for our governance efforts, because of both its (sometime) unpredictability, and our relative lack of understanding about how the systems are successful. AI governance is difficult in large part because we often do not know what an AI will do or why it will behave in a certain way.

In this talk, David Danks will present two different efforts to develop appropriate governance mechanisms for AI systems. First, Professor Danks will discuss efforts to develop a novel framework for dynamic governance, including formal tools to translate legal, societal, ethical, and psychological constraints into constraints on AI performance.

Second, he will examine governance mechanisms for situations that have significant information asymmetries between users/consumers and AI developers/deployers. The goal of these (and other) efforts is to enable better governance of AI systems despite persistent limits on our knowledge.

David Danks is professor of data science, philosophy and policy at the University of California, San Diego. He serves on the National AI Advisory Committee and the Computer Science and Telecommunications Board of the National Academies.

His research lies at the intersection of philosophy, cognitive science, and machine learning, focusing on the ethical and societal impacts of AI and developing novel causal discovery algorithms. Professor Danks is the recipient of a James S. McDonnell Foundation Scholar Award (2008) and an Andrew Carnegie Fellowship (2017). 

This lecture is one of a two-part series. Danks will give a second talk, on Friday, April 25, at 3 p.m., on AI and moral responsibility at the University of Rochester. Travel funding available.

Sponsored by the Central New York Humanities Corridor from an award by the Mellon Foundation.