Policymakers and companies looking to govern advanced artificial intelligence systems are faced with a dilemma: uncertainty. Whether they’re debating the present and future capabilities of the technology, the nature and severity of its risks, or the benefits it might offer, these decision makers are confronted with deep, often unresolvable uncertainty.
In light of this challenge, how should governments regulate AI?
Often, the suggested answer falls near one of two poles: regulatory skepticism (“we don’t know enough; do nothing”) or regulatory prescriptivism (“we know enough; regulate today”). In this talk, Mackenzie Arnold will outline a third option: “radical optionality.” Radical optionality takes seriously the limits of our current knowledge, and the risk that static regulation may quickly become outdated or hinder technological progress. But rather than resolving to do nothing, it proposes a set of actions to maintain flexibility and inform future decision making. By focusing on managing uncertainty rather than ignoring it, radical optionality highlights the value of tools that help governments learn, coordinate, reason, and respond. It offers a potential path toward managing uncertainty.
Speaker
Mackenzie is Director of US Policy at LawAI, where he provides analysis and advice to ensure that advances in AI benefit the public at large. His own research focuses on administrative law, agency decision making, and liability. Prior to joining LawAI, Mackenzie clerked for Judge Joseph A. Greenaway, Jr. of the Third Circuit Court of Appeals, worked in public health law at a New York nonprofit, and graduated, cum laude, from Harvard Law School. Before law school, Mackenzie completed a Fulbright Grant in Ourense, Spain and received his B.A. in political science, summa cum laude, from Boston College, winning the G.F. & J.W. Bemis Award (for exemplary service to others) and the Donald S. Carlisle Award (awarded to the top graduate in political science).
