AI Training Data And Copyright Infringement In India: A Critical Analysis Of ANI V. OpenAI And The Fair Dealing Doctrine
Shaurya Raj, LL.M., National Law University, Jodhpur
Vanshika Arora, LL.M., National Law University, Delhi
1. ABSTRACT
This project critically examines the question whether the “fair dealing” exception under Section 52(1)(a) of the Copyright Act, 1957 can be invoked to justify the use of copyrighted news content to train large language models, as the Delhi High Court provisionally held in ANI Media Pvt. Ltd. v. Open AI OpCo LLC. Drawing on the Court’s interim order, the pending intra-court appeal, and pre-AI fair dealing precedents (Civic Chandran, the DU Photocopy case, India TV v. Yashraj Films, Eastern Book Co. v. Modak), the project situates India’s position against the fair-use jurisprudence of the United States (Thomson Reuters v. ROSS Intelligence, Bartz v. Anthropic, Kadrey v. Meta), the United Kingdom’s legislative restraint on text-and- data-mining reform, and the European Union’s opt-out mechanism, contrasted with Germany’s contrary result in GEMA v. OpenAI. It finds that the Court’s reliance on “private or personal use, including research” to excuse commercial, at-scale training strains a closed, enumerated, 1957-era provision, and concludes with recommendations for reform.
