Gillian Hadfield
Last reviewed
Jun 8, 2026
Sources
12 citations
Review status
Source-backed
Revision
v1 · 1,752 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 8, 2026
Sources
12 citations
Review status
Source-backed
Revision
v1 · 1,752 words
Add missing citations, update stale details, or suggest a clearer explanation.
Gillian Kereldena Hadfield (born July 14, 1961) is a Canadian-American legal scholar and economist whose research sits at the intersection of law, economics, and artificial intelligence. She is known for her work on the design of legal and regulatory institutions, for the 2017 book Rules for a Flat World, and for the "regulatory markets" proposal for governing AI that she developed with Jack Clark. In 2025 she became a Bloomberg Distinguished Professor of AI Alignment and Governance at Johns Hopkins University, having previously served as the founding director of the Schwartz Reisman Institute for Technology and Society at the University of Toronto and as a senior policy adviser at OpenAI. [1][2][3]
Hadfield approaches AI alignment and AI governance as problems of what she calls "normative infrastructure": the systems of rules, norms, and enforcement that allow human societies to cooperate at scale. Her central argument is that aligning AI with human values requires building machines that can read and respond to these normative systems, and that governing AI requires new regulatory institutions rather than only new rules. [2][4]
Hadfield was born in Toronto, Ontario, Canada. She earned a Bachelor of Arts with honours in economics from Queen's University in Kingston, Ontario, in 1983. She then attended Stanford University, where she received a Juris Doctor degree with distinction from Stanford Law School in 1988 and a PhD in economics from Stanford in 1990. Her doctoral work in economics was supervised by the game theorist and later Nobel laureate Paul Milgrom, and she also worked with the economist Kenneth Arrow. Her dissertation examined commitment and the design of long-term contracts, foreshadowing a career-long interest in how incomplete contracts and rules shape behavior. [1]
After law school she clerked for Judge Patricia M. Wald on the United States Court of Appeals for the District of Columbia Circuit, then one of the most influential appellate judges in the country. [1]
Hadfield began her faculty career in 1990 as an assistant professor at the University of California, Berkeley, School of Law. In 1994 she moved to the University of Toronto Faculty of Law, where she was promoted to full professor, and from 1999 to 2001 she concurrently held a global law faculty appointment at New York University School of Law. [1]
In 2001 she joined the University of Southern California, where she spent nearly two decades as the Richard L. and Antoinette Schamoi Kirtland Professor of Law and Professor of Economics at the USC Gould School of Law. During this period she held visiting appointments at Columbia Law School (2008), Harvard Law School (2010), and the University of Chicago Law School (2016). Much of her scholarship in these years applied the tools of law and economics and institutional analysis to the structure of legal systems themselves, including widely cited work on the economic costs of professional control over legal markets and, with Barry R. Weingast, on the "microfoundations of the rule of law." [1]
In 2018 Hadfield returned to the University of Toronto. In 2019 she was named the inaugural Schwartz Reisman Chair in Technology and Society and the founding director of the Schwartz Reisman Institute for Technology and Society, an interdisciplinary research hub created with a major gift from Gerald Schwartz and Heather Reisman. She also became a Canada CIFAR AI Chair affiliated with the Vector Institute for Artificial Intelligence in Toronto, anchoring her growing focus on the governance of advanced AI. [1][5]
The following table summarizes her principal appointments.
| Years | Position | Institution |
|---|---|---|
| 1990 to 1994 | Assistant Professor of Law | University of California, Berkeley |
| 1994 to 2001 | Professor of Law | University of Toronto |
| 2001 to 2018 | Kirtland Professor of Law and Professor of Economics | USC Gould School of Law |
| 2018 to 2023 | Senior Policy Adviser | OpenAI |
| 2019 to 2025 | Founding Director and Schwartz Reisman Chair | Schwartz Reisman Institute, University of Toronto |
| 2025 to present | Bloomberg Distinguished Professor of AI Alignment and Governance | Johns Hopkins University |
Hadfield's best-known book, Rules for a Flat World: Why Humans Invented Law and How to Reinvent It for a Complex Global Economy, was published by Oxford University Press in 2017. Picking up where Thomas Friedman's The World Is Flat left off, the book treats law as a kind of infrastructure, a platform of rules about who can do what, when, and how, and argues that modern legal systems have become too slow, too costly, and too local for a complex global economy. Rather than calling for less law, Hadfield argues for redesigning how rules are produced and delivered, including through more market-oriented and competitive mechanisms. A later paperback edition added a new prologue on artificial intelligence and its regulation. The book built on her earlier edited volume The Second Wave of Law and Economics (1999, with Megan Richardson). [6]
Hadfield's AI work extends her thesis about rules and normative infrastructure to machines. She frames both alignment and governance as questions about how artificial agents can participate in, and be constrained by, the normative systems that humans use to cooperate.
The proposal for which Hadfield is best known in AI policy is "regulatory markets," which she introduced in 2019 with Jack Clark, then a policy leader at OpenAI and later a co-founder of Anthropic. Under the model, set out in the paper "Regulatory Markets for AI Safety," governments do not write detailed technical rules directly. Instead they license private regulators and require the firms being regulated to purchase oversight services from one of them, while holding the private regulators accountable for achieving outcomes set by democratically elected officials. [7]
Hadfield argues that this structure addresses two weaknesses of conventional approaches. Traditional command-and-control regulation struggles with a technical deficit, because legislatures and agencies cannot easily translate broad goals into precise technical specifications for fast-moving technology. Pure industry self-regulation, by contrast, suffers a democratic deficit, because it leaves value-laden choices to companies rather than accountable public actors. By harnessing market competition and private research and development in service of publicly set objectives, regulatory markets are meant to capture the strengths of both. She and collaborators expanded the framework in the 2023 paper "Regulatory Markets: The Future of AI Governance," and the Schwartz Reisman Institute hosted a 2024 workshop to examine how the idea might be operationalized. [7][8][9]
On the technical side, Hadfield has argued that the AI alignment problem closely resembles the economic problem of incomplete contracting. In the 2019 paper "Incomplete Contracting and AI Alignment," written with the computer scientist Dylan Hadfield-Menell and presented at the AAAI/ACM Conference on AI, Ethics, and Society, she contended that humans never fully specify their objectives, whether in contracts or in reward functions, and instead rely on shared norms, courts, and other external institutions to fill the gaps. Aligning AI, on this view, requires giving machines the capacity to interpret and defer to those normative structures rather than optimizing a fixed, literal objective. [10]
This line of work led Hadfield to study the emergence of norms and enforcement directly. In a 2022 paper in the Proceedings of the National Academy of Sciences, she and co-authors used multi-agent reinforcement learning to show that "spurious" or arbitrary norms can help artificial agents learn the more general skills of compliance and enforcement, suggesting that a capacity for normativity may be foundational to cooperative behavior. She also co-authored a 2021 comment in Nature on "cooperative AI," arguing that AI systems must learn to find common ground with humans and with one another. To pursue this agenda she leads the Normativity Lab, a research group focused on building computational models of human normative systems, which moved with her to Johns Hopkins. [1][2]
From 2018 to 2023, alongside her university roles, Hadfield served as a senior policy adviser at OpenAI, where she worked on the governance of frontier AI systems and where the regulatory markets idea took shape. The position placed her among a small group of legal and economic scholars embedded inside leading AI laboratories during the period when large language models moved into mainstream use. [1][7]
In 2025 Hadfield joined Johns Hopkins University as a Bloomberg Distinguished Professor of AI Alignment and Governance, an appointment announced in June 2025. The Bloomberg Distinguished Professorships are interdisciplinary chairs funded by a gift from the philanthropist and former New York City mayor Michael Bloomberg, and Hadfield holds joint appointments in the university's School of Government and Policy and in the Department of Computer Science within the Whiting School of Engineering. She was recruited from the University of Toronto as part of a cluster of hires focused on promoting and governing technological advances. In describing the role, she has said she sees herself as bridging the technical aspects of AI development with the human sciences, and as studying whether AI systems can be given the "normative competence" to operate within democratic institutions. [2][3][11]
Hadfield holds several additional affiliations connected to AI governance and cooperation. She is a senior fellow of the AI2050 program at Schmidt Sciences, board chair of the Cooperative AI Foundation, and a continuing faculty affiliate of the Vector Institute, and she has advised on AI policy through nonprofit initiatives such as Fathom. As of 2026 she remains based at Johns Hopkins, where she continues to publish on alignment, normativity, and the institutional design of AI regulation. [11][12]