Gillian Kereldena Hadfield (born July 14, 1961) is a Canadian economist, legal scholar, and AI researcher. Her work addresses normative social orders, the microfoundations of the rule of law, and AI alignment. She is co-developer, with Barry Weingast, of the Hadfield-Weingast theory of normative social orders, and she proposed regulatory markets as a form of AI regulation while serving as Senior Policy Adviser to OpenAI.
Background and academic career
Hadfield received a BA with honours in economics from Queen's University in 1983, a JD with distinction from Stanford Law School in 1988, and a PhD in economics from Stanford University in 1990 under Paul Milgrom, with Kenneth Arrow as a further adviser. She clerked for Judge Patricia M. Wald of the US Court of Appeals for the District of Columbia Circuit (Source: en.wikipedia.org).
She joined the UC Berkeley School of Law as an assistant professor in 1990, taught at the University of Toronto Law School from 1994 to 2001 and at NYU School of Law's Global Law Faculty from 1999 to 2001, and moved to the USC Gould School of Law in 2001, where she was the Richard L. and Antoinette Schamoi Kirtland Professor of Law and Professor of Economics until 2018 and directed the Southern California Innovation Project and the USC Center in Law, Economics, and Organization. She returned to the University of Toronto in 2018 and in 2019 was appointed Schwartz Reisman Chair in Technology and Society and inaugural director of the Schwartz Reisman Institute for Technology and Society (Source: en.wikipedia.org).
Hadfield was named Bloomberg Distinguished Professor of AI Alignment and Governance at Johns Hopkins University, with joint appointments in the School of Government and Policy and the Department of Computer Science in the Whiting School of Engineering; the appointment was announced by the university in June 2025 (Source: hub.jhu.edu). She is principal investigator of the Normativity Lab, a Canada CIFAR AI Chair at the Vector Institute, and an AI2050 Senior Fellow (Source: en.wikipedia.org).
She served as Senior Policy Adviser to OpenAI from 2018 to 2023, and is a former board member of the American Law and Economics Association and the International Society for New Institutional Economics (Source: en.wikipedia.org).
Regulatory markets
While at OpenAI, Hadfield proposed regulatory markets, in which governments require the targets of regulation to purchase regulatory services from a government-licensed private regulator. The proposal was developed with Jack Clark in "Regulatory Markets: The Future of AI Governance" (arXiv:2304.04914, first posted April 11, 2023, revised February 3, 2026; published in Jurimetrics: The Journal of Law, Science and Technology, vol. 65, pp. 195–240, 2026) (Regulatory Markets: The Future of AI Governance). An earlier and shorter version, "Regulatory Markets for AI Safety," was released as a preprint in December 2019 (arXiv:2001.00078).
The paper identifies two competing problems in AI regulation: a technical deficit, in which legislatures and regulators face difficulty translating command-and-control legal requirements into technical requirements; and a democratic deficit, in which over-reliance on industry to supply technical standards leaves values-based decisions to private rather than democratically accountable actors. Hadfield and Clark argue that regulatory markets could let governments set policy priorities while relying on market forces and industry research to develop the technical methods of regulation (Regulatory Markets: The Future of AI Governance).
The model builds on Hadfield's 2017 book Rules for a Flat World, and the paper develops one application in detail: licensing independent red-teaming and evaluation companies for frontier models, with frontier developers required to contract with a licensed provider (Regulatory Markets: The Future of AI Governance). Hadfield is also a co-author, with Jamie Amarat Sandhu and Noam Kolt, of Regulatory Transformation in the Age of AI (CIFAR, 2023), cited in the same paper for the argument that AI disrupts existing regulatory goals across sectors rather than only creating new harms (Regulatory Markets: The Future of AI Governance).
Democratic matrix and normative competence
Hadfield is the lead author, with Rakshit Trivedi and Dylan Hadfield-Menell, of Building AI for the Democratic Matrix (Knight Columbia, March 3 2026), the paper that develops Normative Competence and the The Democratic Matrix framing (Building AI for the Democratic Matrix: A Technical Research Agenda for Normative Competence and Normative Institutions (Hadfield + Trivedi + Hadfield-Menell, Knight Columbia, March 3 2026)).
The paper proposes normative competence as the technical primitive for democratic alignment of AI agents. Hadfield distinguishes it from preference aggregation, rule encoding, Constitutional AI, "democratic inputs" fine-tuning, and "law-following AI." She argues that each of these implicitly assumes democratic values can be exhaustively elicited and encoded as static parameters, and that this approach will not succeed because democracy is a dynamic complex adaptive system, which she describes as a "dancing landscape" (Building AI for the Democratic Matrix: A Technical Research Agenda for Normative Competence and Normative Institutions (Hadfield + Trivedi + Hadfield-Menell, Knight Columbia, March 3 2026)).
Prior work cited in support of this framework includes:
- Hadfield & Weingast (2012, 2014) — theory of normative social orders, the foundation for normative competence
- Hadfield (2017) — "demand for law" framing
- Hadfield (2024) — original paper introducing normative competence
- Hadfield & Bozovic — informal contracting enforcement
- Hadfield & Ryan — democratic spaces and civic equality
Other publications
Hadfield's work appears in law journals including the Stanford Law Review and in peer-reviewed journals including the Annual Review of Political Science and the Proceedings of the National Academy of Sciences. Selected works include "Problematic Relations: Franchising and the Law of Incomplete Contracts" (Stanford Law Review, 1990), "Legal Barriers to Innovation: The Growing Economic Cost of Professional Control over Corporate Legal Markets" (Stanford Law Review, 2008), "Microfoundations of the Rule of Law" with Barry Weingast (Annual Review of Political Science, 2014), and the book Rules for a Flat World: Why Humans Invented Law and How to Reinvent It for a Complex Global Economy (2017) (Source: en.wikipedia.org).
Her AI-specific work includes "Incomplete Contracting and AI Alignment" with Dylan Hadfield-Menell (AAAI/ACM Conference on AI, Ethics, and Society, 2019), the Nature comment "Cooperative AI: machines must learn to find common ground" with Allan Dafoe, Yoram Bachrach, Eric Horvitz, Kate Larson, and Thore Graepel (2021), and "Spurious normativity enhances learning of compliance and enforcement behavior in artificial agents" (PNAS, 2022) (Source: en.wikipedia.org).
Relationships
- affiliated-with: Vector Institute (Canada CIFAR AI Chair)
- prior-role: OpenAI Senior Policy Advisor (2018–2023); inaugural director, Schwartz Reisman Institute for Technology and Society
- co-author-with: Dylan Hadfield-Menell (son; co-author Knight Columbia 2026), Rakshit Trivedi, Jack Clark (regulatory markets), Barry Weingast (normative social orders framework)
- author-of: Building AI for the Democratic Matrix: A Technical Research Agenda for Normative Competence and Normative Institutions (Hadfield + Trivedi + Hadfield-Menell, Knight Columbia, March 3 2026), Regulatory Markets: The Future of AI Governance
- supports: Normative Competence, The Democratic Matrix, Classification Institutions, Independent Verification Organizations (IVOs)
- engages-with: AI Alignment, Constitutional AI, Law-Following AI
- related: OpenAI