Copyrightability of AI-Generated Works: A Comparative Study of Human Authorship and Creative Control
Abstract
Generative artificial intelligence unsettles copyright's conventional assumption that protectable expression originates in a human author. This paper compares the approaches of the United States, United Kingdom, European Union, China and India to determine when AI-related output can qualify for copyright. Using doctrinal, comparative and normative analysis, it distinguishes AI-assisted works from substantially autonomous outputs and evaluates human authorship, originality, prompts, iteration, selection and post-generation editing. The study finds broad resistance to treating AI itself as an author, but significant divergence in the treatment of human users. The United States emphasizes human-authored elements and partial protection; the United Kingdom retains a statutory rule for computer-generated works; the European Union connects originality with free and creative human choices while regulating AI through transparency duties; China examines individualized prompting and refinement; and India combines a statutory causation rule with an unresolved originality inquiry. The paper argues that copyright should follow identifiable human creative control, with protection limited to human expression, selection, arrangement or modification. Purely autonomous outputs should remain outside ordinary copyright unless legislation creates a narrowly justified alternative right.
How to Cite
Monika, Bhoma Ram Ji, "Copyrightability of AI-Generated Works: A Comparative Study of Human Authorship and Creative Control", Vol. 3, Issue 5, 27-08-2025, pp. 105-115. DOI: 10.5281/zenodo.21350687
