Unjust Enrichment by Algorithm

Abstract

Social media platforms have become enormously powerful, accumulating wealth at an alarming rate, and influencing public opinion with unprecedented efficiency. Platforms use algorithms that promote discriminatory, divisive, extreme, and false content. In recent years, content promoted by social media platforms fueled a series of calamities: the spread of disinformation during the COVID pandemic, the January 6th insurrection, and the establishment of dangerous trends among adolescents and children. The platform crisis is here and is showing no signs of abating. Platform algorithms recommend divisive, hateful and inflammatory content because such content encourages users to spend more time on the platform, allowing platforms to collect more user data, and present users with more advertisements, generating more revenue. Thus, the most socially harmful algorithms are the most profitable for platforms. This profitability is fueling the current crisis: as long as harmful algorithms remain the most profitable, new catastrophes are sure to come.

We argue that any effective legal response to the platform crisis must address the immense profitability of harmful algorithms. We further suggest that this type of legal response is possible through the doctrine of Unjust Enrichment. We explain the conditions under which platform profits should be considered unjust, and how the doctrine of unjust enrichment allows courts to strip platforms of such ill-gotten gains. Rather than prohibit a particular type of content or a specific optimization metric, this proposal targets platforms’ financial incentives, forcing them to consider the broad societal impact of their choices. This is a promising legal venue, offering tools that are unavailable through other frameworks. We detail the advantages of our proposal, explain its origins in existing doctrine of the law of unjust enrichment, and provide a rich account of its implementation in practice.

Keywords: unjust enrichment, social platforms, fake news, disinformation

JEL Classification: K10