Using AI to Mitigate the Employee Misclassification Problem

Abstract

Misclassification of employees as independent contractors is widespread. This article aims to make two contributions. My first goal is to sharpen the explanation of why misclassifications persist; I argue that three well‐known problems – the indeterminacy of employee status tests, the barriers to self‐enforcement, and the inequality of bargaining power – together combine to give employers de facto power to set the default legal status. Putting the burden on the worker to initiate legal proceedings and challenge their classification as an independent contractor is the ultimate reason for persistent misclassifications. The second and main contribution is to propose a solution that relies on new AI capabilities. Thanks to technological advancements it is now possible to require employers to seek pre‐authorisation before engaging with someone as an independent contractor. The authorisation would be granted (or refused) by a state‐run automated system, based on an AI prediction about the law. Both parties would still be able to bring the case before a court of law; but the power to set the default legal status would be taken away from employers. The article considers the difficulties with relying on AI predictions, and argues that those difficulties can be addressed, proposing a model that can be justified.