India’s First AI Training Copyright Case: Court Preliminarily Rules OpenAI’s Use as “Fair Dealing”

In November 2024, Asian News International (ANI) filed a lawsuit against OpenAI before the Delhi High Court. ANI alleged that OpenAI had used its copyright-protected news content without authorization to train the large language model underlying ChatGPT, and claimed that ChatGPT had, in its responses, reproduced or even misattributed ANI’s reporting. ANI also sought an interim injunction, requesting the court to prohibit OpenAI from using or storing ANI’s content during the pendency of the litigation. This case marks India’s first lawsuit involving copyright infringement in the context of generative artificial intelligence.

On July 24, 2026, Justice Amit Bansal of the Delhi High Court rendered his ruling, dismissing ANI’s application for an interim injunction and holding that OpenAI’s use of ANI’s news content to train ChatGPT did not constitute copyright infringement.

The core legal basis for the court’s finding was the “fair dealing” exception under Section 52(1)(a) of the Indian Copyright Act, 1957. This provision permits the use of copyrighted works for research purposes without the need to obtain permission from the rights holder. The court held that OpenAI’s storage and processing of ANI’s news content was, in essence, data analysis for the purpose of training a large language model, and fell within the scope of research-oriented fair dealing, thus not constituting infringement as defined under Section 51 of the Act.

Furthermore, the court noted that OpenAI had used publicly accessible news content and had not circumvented paywalls or subscription barriers. The court observed that if news organizations did not wish their content to be crawled, they could proactively restrict access through network protocols. ANI had failed to establish that ChatGPT had “memorized” or “verbatim reproduced” its news reports in its responses.

Another important factor in the court’s decision to deny the interim injunction was the public interest and industrial development. The judge explicitly stated that granting an interim injunction would seriously impede the development of India’s domestic large language model projects. If training AI required obtaining licenses from every data source, LLM development would become economically unviable. Moreover, an injunction would have widespread negative effects on the public interest, including for the large number of users in India who use ChatGPT free of charge.

It should be noted that this ruling is limited to ANI’s application for an interim injunction; it is a preliminary interim decision at the initial stage and does not constitute a final adjudication on the substantive rights of the parties. The main lawsuit will continue to be tried on its merits. Nonetheless, the ruling carries significant symbolic weight, as it is the first time an Indian court has made a substantive judicial determination on the intersection of generative AI training and copyright law. Against the backdrop of similar disputes still being litigated in multiple jurisdictions worldwide—including the New York Times v. OpenAI case in the United States—Indian courts have taken an initial stance favorable to AI developers at this preliminary stage. The China Intellectual Property Lawyers Network will continue to follow further developments in this case.