OpenAI has secured a major legal win in India after the Delhi High Court said the company’s use of news agency ANI’s content to train ChatGPT did not amount to copyright infringement, marking an important early ruling on how Indian courts may treat artificial intelligence training and copyrighted news material.
Court Says Training Use Was Protected
The ruling centered on whether OpenAI’s storage and use of ANI’s news articles for model training violated India’s Copyright Act.
Reuters reported that Justice Amit Bansal said OpenAI’s storage of ANI’s news articles to train the models underlying ChatGPT was protected by a fair-dealing exemption for research under India’s copyright law.
The decision gives OpenAI an important defense in one of the first Indian lawsuits testing AI training against copyright claims. The remarks was the first substantive court finding in India on whether AI companies can train large language models on copyrighted news content without a license. That makes the case significant beyond ANI because publishers, platforms and AI developers are closely watching how courts define lawful training use.
ANI Failed to Prove Reproduction, Judge Says
The court also addressed whether ChatGPT had copied or reproduced ANI’s work for users. ANI had failed to show that ChatGPT memorized or reproduced its news reports in responses generated for users, according to Justice Bansal. The court’s decision favored OpenAI after ANI accused the U.S. company of using its published content without permission.
That distinction matters because copyright claims often turn on copying, storage, reproduction and output. The court’s finding suggests that simply alleging that copyrighted material was included in training may not be enough if the plaintiff cannot show that the system memorized or reproduced protected material in user-facing responses.
ANI’s Lawsuit Also Raised Fabricated Attribution Claims
ANI’s complaint was not limited to model training. ANI sued OpenAI in the Delhi High Court in November 2024, accusing the company of using its published content without permission and of having ChatGPT attribute fabricated stories to the news agency.
The fabricated-attribution issue reflects a different concern from copyright training: reputational harm when a chatbot incorrectly links false stories to a real news agency. While the court’s finding favored OpenAI on copyright infringement, the broader dispute shows how publishers are challenging AI companies on both content use and the reliability of chatbot outputs.
Ruling Could Shape India’s AI Copyright Debate
The decision lands as courts in other countries are also testing whether AI companies need licenses to train models on copyrighted content. Similar suits against OpenAI are being litigated in the United States and Canada.
For OpenAI, the ruling strengthens the argument that training large language models can fall within lawful research-based use, at least under the facts considered by the Delhi High Court. For news publishers, it shows the difficulty of proving infringement if they cannot demonstrate that an AI system memorized or reproduced protected articles.
The case does not end the global fight between publishers and AI companies, but it gives one of the clearest signals yet from an Indian court. In India, at least for now, using copyrighted news content to train an AI model may be treated differently from copying and republishing that content to users.
