Colloquium: Perspectives on AI Risk
Abstract
In the 18th-century satire Candide, the philosopher Dr. Pangloss remains relentlessly optimistic despite the novella’s stark portrayal of the human condition. His counterpart, Martin, is deeply cynical. Today’s debate over artificial intelligence reflects a similar divide. Some see AI as a Promethean gift to humanity; others warn of existential risks to safety, privacy, and well-being. The truth likely lies somewhere in between—and the intensity of the divide may signal that society has framed the AI problem incorrectly.
In this talk, William Regli offers insights shaped by a decade of public service in science and technology policy. He argues that current AI narratives are overly focused on the technology itself, when the most urgent issues are social in nature. Unlike mature scientific fields such as physics or engineering, we do not yet understand AI’s effects on people and institutions well enough to develop robust governance or engineering practices. As a result, AI’s impact will be uneven—varying in speed, scale, and consequence across different parts of society.
Regli will introduce a practical method for identifying key scientific and engineering questions that arise as AI becomes more deeply integrated with human systems. He will also present examples of research directions worth pursuing. The talk aims to offer a grounded perspective on AI’s trajectory and a tangible approach to navigating its emerging challenges.
About the Speaker
William Regli is a national leader in artificial intelligence, science policy and defense innovation. He is currently a professor of computer science at the University of Maryland, where his research focuses on AI validation, intelligent systems, and digital manufacturing. He has published more than 250 technical articles, holds five foundational U.S. patents in 3D CAD search, and has founded two technology companies.
He recently served in the White House Office of Science and Technology Policy as Senior Advisor for AI Risk, where he supported the National AI Initiatives Office in advancing responsible AI development and deployment across the federal government.
Before his White House appointment, Regli was the founding executive director of the Department of Defense’s Applied Research Laboratory for Intelligence and Security (ARLIS) at the University of Maryland, leading the DoD’s flagship academic center for intelligence and security research from 2018 to 2023. He also held senior leadership roles at the Defense Advanced Research Projects Agency (DARPA), where he served as deputy director and acting director of the Defense Sciences Office from 2014 through 2017. At DARPA, he played a key role in launching cross-cutting programs in AI, sensing, applied mathematics, and advanced manufacturing, and co-developed the agency’s rapid-response Disruptioneering initiative.
His additional government service includes advisory roles with the Department of Energy and the U.S. Air Force Scientific Advisory Board, as well as early-career research at the National Institute of Standards and Technology. He currently serves on the Computing Community Consortium Advisory Board and several editorial boards.
Regli earned his Ph.D. in computer science from the University of Maryland and his B.S. in mathematics from Saint Joseph’s University. He is a Fellow of the IEEE and AAAS, and a senior member of ACM and AAAI.