Human Authorship in AI-Assisted Creative Works: A Three-Factor Test for Evaluating Expressive Control

Abstract

The increasing use of generative Artificial Intelligence in creative production has complicated the determination of human authorship under copyright law. While the United States Copyright Office permits the registration of works containing AI-generated material, the Office currently relies primarily on case-by-case adjudication for such applications, which lacks clear evaluative criteria for determining whether human involvement is sufficient for authorship. This paper argues that a functional test for adjudication is necessary to guide authorship determination. Building on established copyright law and precedents, this paper proposes a functional three-factor test for evaluating human authorship. The proposed test examines expressive control across three stages of creative production: intellectual conception, the generative process, and post-generation modification. By assessing human control over the expressive elements of a work at each stage of creative production, the test provides a structured approach for determining whether the final expressive form of the work can reasonably be attributed to human creative judgment and thereby establish sufficient grounds for human authorship. The proposed framework offers clarity and consistency for copyright adjudication, by providing guidance for better adapting the existing copyright law to the increasing integration of AI technologies in creative production.