Stealing The Style: Conflicts Of Copyright Law In Algorithmic Art
- IJLLR Journal
- 1 hour ago
- 2 min read
Gauri Nandakumar Varma, O.P. Jindal Global Law School
ABSTRACT
Generative artificial intelligence (AI) has presented a new challenge to copyright law by revealing a doctrinal flaw in the treatment of artistic style. Traditionally, copyright did not protect the idea itself and only the expression of an, a doctrine that was formulated in Baker v Selden (1879) and affirmed in the Indian case of RG Anand v Deluxe Films (1978). An artist's style, the patterns of aesthetic choices and composition that characterise a work of art, has been excluded as an unprotectable idea. This was deemed important to avoid monopolies of methods of production and avoid locking up the creative commons. This proposition has been upset by generative AI. Models trained on billions of images can now mimic styles at scale and produce works that are indistinguishable from a particular artist's style, but which do not copy any specific protected work. Case law in Andersen v Stability AI (2023- ongoing) and Getty Images v Stability AI (UK, 2025) highlights the pressing need and existing doctrinal shortcomings. The idea-expression distinction, the doctrine of works derived from other works, the fair use defence and the doctrine of secondary infringement have failed to resolve the problem of the harms to human creators from AI-generated style mimicry. This paper shows that while the doctrine of stylistic exclusion is theoretically valid, its application to generative AI requires fresh consideration. Drawing on case law, the fair use and transformativeness doctrines, the passing off and moral rights doctrines, and the European Union's text and data mining regime, it offers a fair balance through preserving the public domain, while creating specific legal tools, such as transparency, meaningful opt-out, and a brief sui generis protection, to prevent the unconsented commercial use of identifiable styles through generative AI.
