Fairer Public Benefit in Copyright Law

Authors: Amanda Levendowski

Abstract:

In 1966, a court considered expressly whether a secondary use of copyrighted works served a public benefit. While public benefit has become a subfactor of the fair use doctrine, it remains undefined, uncodified, and undertheorized. After the recent Supreme Court decision in Google v. Oracle, however, it is also unavoidable: the Court stated that “we must take into account the public benefits the copying will likely produce.” Previously, courts invoked public benefit with some predictability in pivotal cases involving novel technologies, from home video recorders to digital libraries to algorithms. A hand-coded dataset of nineteen U.S. technology-related public benefit cases from 1966-2023 reveals five values that emerge from those cases: expression, knowledge, entertainment, competition, and/or efficiency. Forthcoming judicial decisions about the latest novel technology, artificial intelligence (AI), will be shaped by this precedent. However, a series of mid-aughts decisions about algorithms exposed an FU long lurking in fair use: name aside, there is nothing particularly fair about it. Those cases excused invasive, coercive, and biased AI systems as efficient “public benefits” when finding them to be fair use. Many scholars have written about the unfairness of fair use, and this article contributes to those conversations by using a feminist cyberlaw lens to critique the practice of dubbing technologies public benefits without acknowledging, let alone assessing, countervailing public harms. A public benefit that ignores public harm is incomplete. Purported fair uses, particularly those underpinning AI systems, can amplify bias, dis/misinformation, and environmental destruction -harms that are predictable, preventable, and passed over by public benefit presently. This article responds by recalibrating public benefits to better account for these and other public harms. It defines a fairer public benefit and develops a framework for realizing it. The latter poses challenges. In courts, public harm has already happened when matters are litigated, placing a premium on compensation rather than prevention. Congress could codify public benefit, but it is unlikely that Congress could agree upon a satisfactory definition. To further complicate matters, neither judges nor legislators have duties of sociotechnical competency. But lawyers do. Client-centered counseling could facilitate a fairer public benefit if there were a framework for doing so. This article proposes one: FAIRR (pronounced “fairer”), a mnemonic for formalizing purposes, assessing benefits, identifying harms, reconsidering those benefits in light of those harms, and reporting to the client. Inspired by computer science’s threat modeling methodology, FAIRR represents a rigorous, repeatable method for analyzing how infringement liability, public perception, and social progress are affected by public benefits and public harms. By deconstructing the inequities embedded in public benefit as they exist now and developing a fairer alternative for the future, this article helps lawyers shape better technologies.

Keywords: intellectual property, copyright, fair use, public benefit, technology, artificial intelligence, bias, disinformation, environmental destruction