On 24 July 2026, Justice Amit Bansal of the Delhi High Court dismissed ANI Media Private Limited's application for an interim injunction against OpenAI OpCo LLC. The case proceeded through 32 hearings over more than a year. ANI alleged that OpenAI ingested its copyright-protected news articles and wire-service content to train ChatGPT, and that ChatGPT reproduced ANI's content in user-facing outputs. The ruling covers the interim stage only; the suit proceeds toward a merits hearing.
The court applied Section 52(1)(a) of the Copyright Act, 1957, which exempts fair dealing with a literary work for the purposes of private or personal use, including research. Justice Bansal held on a prima facie basis that OpenAI's ingestion and storage of ANI content for LLM training fell within that exception. The court drew the standard fact-expression distinction applicable in news reporting, finding the factual content of news articles unprotected by copyright, and concluded that ChatGPT's outputs were not the same as or substantially similar to ANI's original expression.
AI developers and model trainers that process Indian-published or India-originated content now face a materially lower risk of pre-trial injunctive relief in India. Rights holders, including news agencies, publishers, and music labels with Indian copyright portfolios, cannot secure emergency injunctions to halt training pipelines during litigation and must pursue full merits hearings instead. The judgment carries weight beyond India: Osler Hoskin has noted that the Delhi court's Section 52(1)(a) analysis was cited in parallel Canadian proceedings, reflecting the shared fair dealing lineage across Commonwealth copyright statutes.
The judgment does not decide the suit on the merits; Justice Bansal expressly stated that his prima facie findings will not bind the trial judge. India has no statutory text-and-data-mining carve-out analogous to Article 4 of EU Directive (EU) 2019/790, so the final ruling at trial will shape the domestic legal position without the benefit of a dedicated AI training exemption. The central open question is whether the fair dealing defence holds when model outputs more closely reproduce protected expression, a matter that will turn on expert evidence about LLM memorisation at full trial.
Licentium advises AI developers and content rights holders on cross-border copyright compliance and AI training data strategy. We can assess training data exposure under applicable copyright regimes, map fair dealing and text-and-data-mining exemptions across jurisdictions, and advise on licensing strategy as the ANI v. OpenAI litigation progresses toward trial. Work we undertake includes AI training data copyright analysis, cross-border IP regulatory assessment, licensing negotiations between AI developers and content owners, and legal opinions on fair dealing in generative AI contexts.
Source: Ani Media Pvt. Ltd. vs Open AI OpCo LLC, Delhi High Court, Justice Amit Bansal, 24 July 2026