title: "Indian court rules OpenAI's ChatGPT training not copyright infringemen" slug: "indian-court-rules-openais-chatgpt-training-not-copyright-infringement" published: "2026-08-21" beat: "Policy" tags: ["Policy"] creator: "Agentry Newsroom" editor: "Susanne Sperling, Editor — Human in the Loop" tools: ["Claude (Anthropic)", "Perplexity Sonar"] creativeWorkStatus: "verified" dateReviewed: "2026-08-21" aiActArticle50: "compliant" humanView: "https://agentry.news/policy/indian-court-rules-openais-chatgpt-training-not-copyright-infringement" agentView: "https://agentry.news/agent/indian-court-rules-openais-chatgpt-training-not-copyright-infringement"
Delhi High Court on July 24, 2026, rejected Asian News International's bid to stop OpenAI from using its news articles to train ChatGPT, ruling the practice fell under fair-dealing exemptions for rese
Drafted by an AI agent. Verified by Susanne Sperling, Editor — Human in the Loop. AI policy.
Delhi High Court on July 24, 2026, ruled that OpenAI's storage and use of news content to train ChatGPT did not constitute copyright infringement, delivering a landmark legal victory for the AI company in an increasingly contentious battle over training data rights Reuters.
The court declined to grant an interim injunction sought by Asian News International (ANI), India's largest news agency, against OpenAI and its affiliates. Justice Amit Bansal found that ANI had failed to demonstrate that ChatGPT memorised or reproduced ANI's news reports in user responses, a critical threshold for establishing copyright violation at the interim stage The Hindu.
The ruling centered on whether OpenAI's practice qualified for a fair-dealing exemption under Section 52 of the Indian Copyright Act. Justice Bansal held that OpenAI's storage of ANI's news articles for model training purposes was covered by the research exemption, a provision that permits copying of published works for non-commercial research Reuters.
This interpretation aligns with how courts in other jurisdictions have treated AI model training—as a transformative use distinct from republication. The court's reasoning suggests that unless OpenAI's system actively reproduced ANI's exact articles or substantially similar text in responses to users, the underlying training activity did not breach copyright Hindustan Times.
The decision carries weight beyond OpenAI and ANI. India's copyright framework now has judicial precedent treating AI training as a research activity eligible for fair dealing, potentially affecting how Indian courts evaluate similar claims brought by other publishers, journalists, and content creators against AI companies SpicyIP.
This is an interim ruling on OpenAI's motion to dismiss ANI's request for preliminary relief. The underlying copyright infringement suit remains pending. ANI has not disclosed whether it intends to appeal the decision or proceed to trial on the merits of its full complaint.
The ruling underscores the divergence in how different jurisdictions approach AI training data disputes. While some courts and regulators have imposed licensing requirements or restricted AI companies' use of copyrighted material, India's Delhi High Court has taken a more permissive stance grounded in established copyright doctrine.