The use of AI in legal systems: determining independent contractor vs. employee status

Maxime C Cohen; Samuel Dahan; Warut Khern-Am-Nuai; Hajime Shimao; Jonathan Touboul

doi:10.1007/s10506-023-09353-y

The use of AI in legal systems: determining independent contractor vs. employee status

Artif Intell Law (Dordr). 2023 Mar 30:1-30. doi: 10.1007/s10506-023-09353-y. Online ahead of print.

Authors

Maxime C Cohen^#¹, Samuel Dahan^#², Warut Khern-Am-Nuai^#¹, Hajime Shimao^#², Jonathan Touboul^#³

Affiliations

¹ Desautels Faculty of Management, McGill University, 1001 Rue Sherbrooke O., Montreal, QC H3A 1G5 Canada.
² Faculty of Law, Queen's University, 128 Union St, Kingston, ON K7L 2P1 Canada.
³ Department of Mathematics, Brandeis University, 415 South Street, Waltham, MA 02453 USA.

^# Contributed equally.

Abstract

The use of artificial intelligence (AI) to aid legal decision making has become prominent. This paper investigates the use of AI in a critical issue in employment law, the determination of a worker's status-employee vs. independent contractor-in two common law countries (the U.S. and Canada). This legal question has been a contentious labor issue insofar as independent contractors are not eligible for the same benefits as employees. It has become an important societal issue due to the ubiquity of the gig economy and the recent disruptions in employment arrangements. To address this problem, we collected, annotated, and structured the data for all Canadian and Californian court cases related to this legal question between 2002 and 2021, resulting in 538 Canadian cases and 217 U.S. cases. In contrast to legal literature focusing on complex and correlated characteristics of the employment relationship, our statistical analyses of the data show very strong correlations between the worker's status and a small subset of quantifiable characteristics of the employment relationship. In fact, despite the variety of situations in the case law, we show that simple, off-the-shelf AI models classify the cases with an out-of-sample accuracy of more than 90%. Interestingly, the analysis of misclassified cases reveals consistent misclassification patterns by most algorithms. Legal analyses of these cases led us to identify how equity is ensured by judges in ambiguous situations. Finally, our findings have practical implications for access to legal advice and justice. We deployed our AI model via the open-access platform, https://MyOpenCourt.org/, to help users answer employment legal questions. This platform has already assisted many Canadian users, and we hope it will help democratize access to legal advice to large crowds.

Keywords: Artificial intelligence; Decision-support legal systems; Employment law; Predictive models.

© The Author(s), under exclusive licence to Springer Nature B.V. 2023, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.