Courts and regulators have increasingly found that AI-generated content is not copyrightable, but that works created with AI assistance may receive protection for the human author’s contribution to or arrangement of AI-generated elements. This project identifies an unforeseen consequence of this framework: it transforms copyright infringement analysis from a content-based inquiry into a process-based one. Because AI-generated content is often indistinguishable from human-authored material on its face, courts cannot perform the “filtration” analyses that copyright litigation typically requires—separating protectable from unprotectable elements—without investigating the process by which a work was created. Unlike ideas, facts, or other public domain materials, which can be identified from a work’s content, the scope of copyright in an AI-assisted work can be identified only through examination of the prompts, outputs, iterations, and editing decisions that gave rise to the final work. This shift has profound consequences for copyright’s litigation system: early disposition will be all but impossible, discovery costs will balloon, and opportunities for strategic behavior will abound.
The project identifies this problem at its inception and considers solutions that might head it off. Courts might be tempted to abandon rigorous comparison in favor of holistic analysis or an identicalness standard, but both of these approaches would undermine copyright’s policy agenda. Instead, the project proposes a range of mechanisms that would allow courts to continue using filtration, but make it more administrable: a graduated evidence disclosure requirement in litigation, under which plaintiffs who fail to produce creation records would face a heightened similarity standard; an expanded evidentiary role for the Copyright Office as a repository for AI process documentation; and technological tools that can help verify and refine the evidentiary record. Together, these proposals aim to preserve the most analytically sound form of infringement analysis while adapting its evidentiary demands to the realities of a world of ubiquitous AI-assisted creation.
Jacob Noti‑Victor is a professor of law. He teaches and writes in the areas of intellectual property, technology law, and property. His scholarship focuses on how legal regimes shape innovation, creativity, and the adoption of emerging technologies, including artificial intelligence. Noti‑Victor's work has appeared in leading law reviews such as the Virginia Law Review, Stanford Law Review, Minnesota Law Review, and Washington University Law Review. His current research explores issues such as AI authorship, competition in copyright markets, and press freedom. Before entering academia, Noti‑Victor practiced intellectual property litigation at Kirkland & Ellis LLP and clerked for Judge Pierre N. Leval on the U.S. Court of Appeals for the Second Circuit. He holds a J.D. from Yale Law School, where he was an Essays Editor of the Yale Law Journal, and an A.B. from Harvard College.
Sponsoring Organization(s)
Information Society Project