In the first Indian decision on whether training a large language model on copyrighted material infringes copyright, Justice Amit Bansal of the Delhi High Court denied ANI Media's plea for an interim injunction against OpenAI on July 24, 2026. ANI raised both an output claim and an input claim, arguing this fell within Section 14(a)(i) of the Copyright Act, 1957. OpenAI relied on the fair dealing exception for research under Section 52(1)(a), arguing that training is transformative. OpenAI won on both counts. The Court held that although the Copyright Act is not extraterritorial, training on ANI's corpus and resulting outputs formed one continuous process, and since ANI's principal place of business was in India and outputs were accessed here, the Act applied. ANI's exhibits showed ChatGPT producing answers nearly identical to its articles, but the models named in the suit, GPT-4 and GPT-4o, were trained before the articles were published. The Court stated that “the illustrations given in the plaint are post the training of Open AI's LLMs and a case for memorization… cannot be made out” (para 124, page 60). The Court left open whether displaying content through live retrieval amounts to “communication to the public” under Section 2(ff) of the Copyright Act, noting the issue but not deciding it since ANI had not properly pleaded it.