Copyright Authorship in the Era of Generative AI: A Control-Based Framework

Abstract

Generative artificial intelligence (AI) fundamentally challenges the traditional conception of authorship within copyright law. It was long thought that only human creativity can generate copyrightable works. Recent advances enable the production of expressive content with minimal or no direct human input, raising complex questions regarding originality, creative contribution, and attribution. In many places, courts and lawmakers have mostly agreed that people should be the ones who write things. However, the swift expansion of AI-assisted and AI-generated works has exposed significant doctrinal ambiguity regarding originality, creative contribution, and attribution. This article examines copyright law's treatment of authorship in the era of generative AI. Using doctrinal analysis, comparative scholarship, and institutional guidance, it looks at different ways to use AI to create things. These include strict human-centered authorship models, partial or hybrid human–AI authorship theories, and institutional responses based on policy that try to fix problems with incentives and the market. The emphasis is on the evolving nucleus of creativity within generative systems, where human contribution increasingly manifests through prompt design, iterative refinement, selection, and the establishment of constraints, as opposed to conventional expressive execution. The Article proposes a control-based framework for copyright authorship, arguing that substantial human creative control, rather than mere mechanical expression, should be the primary criterion for authorship attribution in generative AI-mediated creation. This method keeps the human-centered roots of copyright while making authorship law more relevant to how people make things today.