Release Note
Fair Use watch, 2026 Issue
Fair use is ‘a right’ to use an otherwise ‘protected’ work. In its most general sense, a fair use is any copying of copyrighted material done for a limited and “transformative” purpose, such as to comment upon, criticize, or parody a copyrighted work. Such uses can be done without permission from the copyright owner. In other words, fair use is a defense against a claim of copyright infringement. If your use qualifies as a fair use, then it would not be considered an infringement.[i]
Imagine a Generative Artificial Intelligence (AI) Software ‘ZHK’, that claims it can conduct research and generate pictures with accuracy. How is it able to do it? AI Systems (including ZHK) work on ‘heuristics’, picking up patters in the data they are trained on. Here ZHK’s training module may consist of innumerable books, researches, review papers, paintings, photographs etc. on multiple subjects based on which its responses sound intelligent, albeit artificially. While no author/researcher/artist/photographer was paid any royalty or consented to this arrangement, ZHK still has access to these sources. Based on them, ZHK now responds to queries, often generating sub-standard responses that ‘sound’ or ‘look’ similar but are devoid of any literary or creative nuance. Unfettered by any paywall or subscription fees or even the decency of giving credits to the author, ZHK allows any user to access ‘data’ in its training corpus. Now the question here is: Does this unlicensed appropriation and blatant disenfranchisement of the creator’s right constitute fair use? Courts around the world have been confronted with the research question for this repository in some or the other way. The question being, the scope of the application of the Doctrine of Fair Use in relation to the training of Large Language Models (LLMs) and Generative AI systems.
The necessity of this repository arises from the unresolved legal and ethical challenges posed by generative AI. Although courts have acknowledged the risks posed by generative AI, such as the displacement of human creativity and the erosion of incentives for authors, they have often stopped short of providing clear regulatory principles. Judgments frequently hinge on evidentiary shortcomings rather than substantive rulings, leaving critical questions unresolved. This repository therefore serves as both a scholarly resource and a policy-relevant guide, offering a structured analysis of statutory provisions, case briefs, and judicial reasoning, while also pointing to the gaps that demand future legislative and judicial attention.
This repository is the product of sustained research into the evolving relationship between copyright law and artificial intelligence, with a particular focus on the fair use doctrine as applied to the training of Large Language Models (LLMs). It brings together statutory foundations, judicial precedents and critical commentary to provide a consolidated repository of knowledge on how courts having common law jurisdiction have grappled with the legality of using copyrighted works in AI training.
While the United States has produced the most visible case law on the subject, the repository also examines parallel developments in Germany, the United Kingdom, France, India, Canada, and the European Union. Together, these jurisdictions reveal both convergences and divergences in how courts and legislatures are responding to the challenges posed by training Large Language Models (LLMs) on copyrighted works.
By examining provisions of the Copyright Act of 1976, the Digital Millennium Copyright Act, the Lanham Act, Urheberrechtsgesetz (UrhG), InfoSoc Directive and Indian Copyright Act 1957 alongside landmark cases such as Kadrey v. Meta, Bartz v. Anthropic, Raw Story Media v. OpenAI, Thomson Reuters v. Ross Intelligence, ANI Media Pvt. Ltd. v. OpenAI, CANLII V. CasewayAI, the repository offers a comprehensive view of the doctrinal tensions between innovation and the protection of authors’ rights.
The aim of this repository is to critically assess whether the use of copyrighted works in AI training can be justified under the fair use doctrine, and to highlight the inconsistencies and unanswered questions that remain in judicial reasoning. It seeks to provide clarity on how courts apply the four statutory factors of fair use, i.e., purpose and character, nature of the work, amount and substantiality, and market effect, when confronted with disputes involving generative AI. The objective is not only to document existing case law but also to underscore the broader implications of these rulings for authors, AI developers, and society at large.
Structure of the Report
The entire report follows a uniform sequence, for the better understanding of the reader, the sequence has been reproduced below:
Jurisdiction
Laws Applicable
Cases and Analysis
Approach of the Courts
Disclaimer: For certain jurisdictions, the statutes perused were translated to English Language. Diligent reliance was placed upon translation tools like Google Translate. The output was then scrutinised to the best of our understanding.
Discrepancies, if any, maybe reported to the project lead Halata Zehra at halatazehra@virtuositylegal.com.
[i] Stim R at, ‘What Is Fair Use?’ (Stanford Copyright and Fair Use Center, 25 November 2021) <https://fairuse.stanford.edu/overview/fair-use/what-is-fair-use/> accessed 15 March 2026
Leave a Reply