The Authorship Dilemma: Navigating Copyright Ownership in the Age of Generative AI
TL;DR
The answer lies not in deciding whether a machine can be an author but in reforming copyright to separate authorship, creative contribution and ownership, and a tiered framework calibrated to the degree of human creative control is proposed.
Abstract
Copyright law rests on the premise that a protected work originates in the intellectual labour of an identifiable human author, and it allocates ownership, economic rights and control by reference to that authorship. Generative artificial intelligence unsettles that premise by producing expressive work through algorithmic processes in which the human contribution may be confined to a prompt, a selection among outputs or a subsequent refinement. This article examines the resulting dilemma. It shows that the traditional doctrines of originality, human authorship, the idea and expression dichotomy, ownership, work made for hire, joint authorship and derivative works yield inconsistent or indeterminate results when applied to machine-generated material, and that the difficulty is one of incompatibility rather than of interpretation. It maps the possible claimants to authorship, being the user, the developer, the platform owner and the contributors of training data, and finds that none satisfies the classical test, so that the public domain becomes a plausible default. It compares the positions taken in the United States, the United Kingdom, the European Union, India, China and Australia, and considers the philosophical theories of personality, labour and utility. It concludes that the answer lies not in deciding whether a machine can be an author but in reforming copyright to separate authorship, creative contribution and ownership, and proposes a tiered framework calibrated to the degree of human creative control.
