Fair Use by Design
19 Pages Posted: 8 Aug 2018
Date Written: 2017
Abstract
Copyright law seeks to promote the creation of works for the benefit of the public. Fair use doctrine was intended to serve as a check on copyright, to ensure that the law achieves its objectives. Fair use thus limits the rights of authors, and legitimizes unlicensed use of copyrighted works, whenever a rigid application of rights would prevent socially beneficial uses and possibly conflict with the public interest. Nowadays, the vast majority of copyrighted materials are distributed digitally, and copyright is enforced algorithmically. Fair use, as a legal defense against infringement allegations might be largely irrelevant where algorithmic enforcement keeps disputes out of court. In this paper I argue that algorithmic adjudication requires a new approach to fair use. In a nutshell, I argue that in order to serve its role in this era, fair use must be embedded in the system design. Fair use by design seeks to apply artificial intelligence and machine learning, to identify potential non-infringing uses. Algorithms could learn to identify patterns of fair use instances by studying previously decided fair use cases. The system will flag instances where fair use is highly probable, and may transfer inconclusive cases to human review, or subsequently to courts. Machine learning capabilities will ensure that the system incorporates new fair use rulings. Fair use by design is not simply a technical patch. Aside from the technological challenges, it also involves legal innovation. It requires a new framework for addressing algorithmic governance which involves redefining the role of courts and the nature of judicial oversight. The paper demonstrates these new legal challenges and discuss their implications for fair use doctrine. Part I begins by briefly describing the rise of algorithmic enforcement in copyright, where access to copyrighted materials, and copyright disputes, are increasingly governed by algorithms. Part II explains why this may put fair use and freedom of speech in danger. In Part III, I introduce the notion of fair use by design and explain how this approach can help address some of these threats. Finally, Part IV outlines some of the legal challenges that I anticipate in making this transition to algorithmic adjudication.
Keywords: fair use, algorithmic, by design, code as law, copyright, user rights, artificial intelligence, machine learning
JEL Classification: K00
Suggested Citation: Suggested Citation