Portrait of Ignacio Cofone

Ignacio Cofone

Affiliate Member
Professor, University of Oxford
Research Topics
AI Alignment
AI Ethics
AI Legal Issues
AI4Humanity
Algorithmic Fairness
Human-Centered AI
Law
Privacy
Trustworthy AI
Website
Google Scholar
LinkedIn

Biography

Ignacio Cofone is the Professor of Law and Regulation of AI at Oxford, appointed jointly at the Institute for Ethics in AI and the Faculty of Law. He is also an Affiliated Fellow of the Yale Information Society Project.

Ignacio’s research examines how AI changes legal concepts such as fairness, privacy, and responsibility, and how law should govern the risks and harms that AI creates through inference and prediction. He is currently writing Plausible (Cambridge University Press), on how to govern reliance on AI systems.

His previous book, The Privacy Fallacy (Cambridge University Press 2023), argues that AI harms require privacy and data protection law to rest on duties of non-exploitation, because individual control over data has become ineffective. Before joining Oxford, he held the Canada Research Chair in AI Law and Data Governance at McGill University.

Publications

The Overstated Cost of AI Fairness in Criminal Justice
Ignacio Cofone
Warut Khern-am-nuai
A dominant critique of algorithmic fairness holds that increasing fairness reduces predictive accuracy, imposing a cost on society. We chall… (see more)enge that assumption by empirically analyzing the COMPAS dataset. We make two contributions. First, using causal inference methods, we show that racial bias is not only present in the COMPAS dataset but is also amplified by the models trained on it. Widely used models do more than replicate existing bias; they exacerbate it. This undercuts both the assumption that algorithmic decision-making offers a neutral improvement over human judgment and the weaker claim that it merely mirrors preexisting human bias. Second, we reframe the fairness-accuracy tradeoff. Applying fairness constraints does not necessarily cost predictive accuracy in criminal justice. Prediction systems operationalize concepts such as risk through implicit and often flawed normative choices about what to predict and how. The tradeoff claim assumes that the unconstrained model's prediction is an optimal baseline. Fairness constraints can instead correct distortions introduced by biased outcome variables: rearrest data, in this case, captures and magnifies systemic racial disparities. Under some interventions, therefore, fairness carries none of the cost presumed in policy debates. These dynamics extend beyond criminal justice to lending, hiring, and housing, where biased outcome variables reinforce inequality independently of proxy selection. We draw out what this implies for how law and policy should approach fairness adjustments in criminal law.
2026-07-31
ArXiv (preprint)
doi.org
arxiv.org