Delhi High Court Refuses Interim Injunction Against OpenAI in ANI Copyright Case

The Delhi High Court on Friday, July 24, 2026, declined to grant interim relief to news agency Asian News International (ANI) in a copyright infringement suit against OpenAI. The court held, prima facie, that OpenAI's use of ANI's works for training its ChatGPT chatbot falls within the exception under Section 52(1)(a) of the Copyright Act.

Justice Amit Bansal ruled that storing ANI's literary works for training Large Language Models (LLMs) did not amount to copyright infringement at this preliminary stage. The court also found that responses generated by ChatGPT using Retrieval-Augmented Generation (RAG) were not substantially similar to ANI's original works.

Additionally, the court observed that restraining OpenAI at this stage would cause irreparable injury not only to the company but also to the public at large. This decision comes in the first copyright action by an Indian news organisation against OpenAI, filed by ANI in November 2024.

ANI had alleged that OpenAI used its copyrighted news reports without permission to train ChatGPT, and that the chatbot generated responses that copied or closely summarised its reports. ANI also claimed that ChatGPT falsely attributed fabricated news reports and statements to the agency, harming its reputation.

According to court records, ANI offered to grant a license to OpenAI for using its news content for LLM training on October 3, 2024, but the US-based company declined the proposal. During the hearing, ANI's counsel stated that OpenAI had entered into strategic partnerships and licensing arrangements with other news organisations such as the Financial Times, Associated Press, and Condé Nast, which operate on a similar business model to ANI.

OpenAI denied the allegations, maintaining that ANI's website had been excluded from future training and that ChatGPT did not unlawfully reproduce ANI's works.

This ruling is significant as it addresses the intersection of copyright law and artificial intelligence training in India. Section 52(1)(a) of the Copyright Act permits the use of copyrighted works for the purpose of research or private study, or for criticism or review. The court's prima facie finding suggests that using copyrighted materials to train AI models may be considered fair dealing under Indian law, though the final decision awaits a full trial.

The case underscores the growing global debate over whether AI companies should pay for the data they use to train their models. While some news organisations have entered into licensing agreements with OpenAI, others, like ANI, have chosen to litigate. The Delhi High Court's decision is likely to influence future cases on AI and copyright in India.