The Berkman Klein Center community welcomes former executive director Urs Gasser back to Harvard to celebrate the launch of his new book, Guardrails: Guiding Human Decisions in the Age of AI in conversation with Lis Sylvan and Bruce Schneier. For twenty years, Dr. Gasser has been a steward and champion of BKC, including a twelve-year tenure as executive director. During that time, he guided the center through a substantial period of growth and brought a human-centric lens to the shifting sands of a dynamic technological world.
This characteristic lens acts as a guiding force in Dr. Gasser’s latest book, Guardrails, which argues that in an age of computer-assisted decision making, the most-needed innovation is not technological, but social.
Drawing on more than two decades of research, Guardrails offers a new framework for navigating human decision making, one that embraces individual agency while remaining aware of the influence of social context.
In Guardrails, Gasser and his co-author, Internet governance expert Viktor Mayer‑Schönberger, invite the reader to critically examine the societal “guardrails” that guide decision-making – the constraints on the information available to us, the norms that govern how we choose among our options, and the consequences that bind us to our choices. To address the concerns of our time, we must understand the scope of guardrails available to us and craft guardrails appropriate to the digital age.
About Guardrails:
As we progress in the digital age, the locus of power and thus our focus of regulation changes. It used to be the network: Who has what kind of connectivity, who controls the network's bottlenecks including the ability to store and process information, which interactions will the network allow to take place, and which will it retard, reduce, and restrict? What virtual spaces and platforms would emerge, and how would they exercise and project power? More recently, our focus turned to data: Who can collect and utilize data? What patterns do massive amounts of data reveal that smaller sets hide, and who gets to see these insights? What will machines learn from data? Accordingly, data governance and the geopolitics of information have become hot topics.
In Guardrails, Dr. Gasser and his co-author, Internet governance expert Viktor Mayer‑Schönberger, argue that in the age of AI, a focus on data is mistaken, or at least dangerously incomplete. Because AI's real promise isn't better access to insights; it is improved decisions. The more we realize humans' cognitive flaws and experience the power of data-driven machine learning, the more we are tempted to delegate decision agency to the machine. What we need to focus on is no longer network governance, nor data governance, it's decision governance. And for this focus, we are conceptually ill prepared. To jump-start the discussion, Guardrails puts forward four design principles for decision governance and suggests that the real innovation we need to make it happen is less technical than social.
About the author:
Urs Gasser is Professor of Public Policy, Governance, and Innovative Technology at the Technical University of Munich, where he serves as Dean of the newly launched TUM School of Social Sciences and Technology, and Rector of the Munich School of Politics and Public Policy. Before his appointments in Munich, Urs was a Professor of Practice at Harvard Law School and, from 2009 to 2021, Executive Director of the Berkman Klein Center for Internet & Society, where he remains a member of the board of directors.