Technological developments often serve as catalysts for new fields of study. Powered by data and learning techniques, algorithms are now implementing balances among societal values that were traditionally addressed by legal doctrines. This development has already motivated successful research on differential privacy and more recently algorithmic fairness. Is it time for a broader research effort on algorithms and law? The goal of this talk is to facilitate a discussion on what a joint research field involving algorithm designers and legal scholars could possibly look like. We focus in particular on algorithm design rather than computer science as whole, leaving “computational law” topics such as legal text mining out of scope for this talk.

At a very high level, an algorithm is not far from a legal doctrine: an algorithm gets input and makes a decision, with the goal of optimizing some objective subject to constraints; similarly, a legal doctrine is applied to decide cases by balancing among different objectives and constraints. We cover two recent examples of algorithms required to implement legal doctrines (beyond privacy and fairness): (1) Autonomous driving and duty of care; (2) Algorithmic pricing and antitrust. In the opposite direction, we discuss how personalization of law is bringing legal doctrines closer to algorithms that take input details into account. We close by raising perhaps the biggest challenge to an interdisciplinary research field of algorithms and law – bridging between the mathematical and text-based approaches.