Indian court rules OpenAI did not infringe ANI copyright
The Delhi High Court rejected a copyright infringement claim against OpenAI on July 24, 2026, ruling that the company's use of news agency ANI's articles to train ChatGPT fell within India's fair-dealing exemption for research. Justice Amit Bansal, hearing the case of Asian News International v. OpenAI OpCo LLC, declined ANI's application for an interim injunction, finding that OpenAI's storage of the news articles did not amount to copyright infringement at the preliminary stage Reuters.
The Ruling and Fair-Dealing Defense
Justice Bansal's decision centered on Section 52(1)(a) of India's Copyright Act, which permits use of copyrighted material for research purposes. The court found that OpenAI's incorporation of ANI content into its training dataset qualified as protected research activity The Hindu. Critically, ANI failed to demonstrate that ChatGPT had memorized or reproduced ANI's original news reports in its responses to users—a factual gap that weakened the agency's claim of direct copyright harm Reuters.
The interim-injunction phase of this case did not resolve the underlying copyright suit, meaning the parties may pursue further litigation. However, the court's ruling that training-related storage constitutes fair dealing for research establishes a significant legal precedent for AI model developers operating in India.
Implications for the AI Agent Economy
This decision is material for companies building large language models and AI agents that rely on web-scraped training data. By endorsing the fair-dealing exemption for research, the Delhi High Court has signaled that Indian copyright law permits data ingestion for AI training without requiring individual licenses from content creators Economic Times. This contrasts with emerging legal frameworks in other jurisdictions where content licensing and consent requirements are being litigated or mandated.
No penalties, fines, or regulatory action against OpenAI resulted from this ruling, as it addressed only the threshold question of whether an interim injunction was warranted. The court found no prima facie case of infringement strong enough to suspend OpenAI's operations pending trial.
The case underscores the ongoing tension between publishers and AI developers over training data rights, a friction point that will likely define the agent economy's compliance landscape across multiple geographies.