ANI takes OpenAI copyright battle to Delhi HC division bench

MUMBAI: The copyright battle between news agency Asian News International (ANI) and OpenAI is heading into another round, with ANI challenging a Delhi High Court single judge’s decision to reject its plea for interim protection against the use of its content by ChatGPT.

ANI filed an appeal before a division bench of the Delhi High Court on Tuesday. The matter was listed before a bench headed by Justice V Kameswar Rao, but the bench did not assemble.

Justice Rao has since been appointed Chief Justice of the Patna High Court. The appeal has now been listed for hearing on September 14.

The case, titled ANI Media Pvt. Ltd. v. OpenAI Inc & Anr., stems from ANI’s copyright infringement suit against OpenAI, in which the news agency has alleged that its original news content is being used by ChatGPT for commercial gain.

ANI had argued before the single judge that ChatGPT reproduces its original news reports verbatim when responding to users’ queries in real time.

The agency also alleged that the chatbot had attributed statements and news reports to ANI that had never actually been published or occurred, raising concerns over both copyright and the accuracy of AI-generated responses.

The dispute is significant because it concerns the use of copyrighted news material to develop and operate large language models, as well as the extent to which AI systems can reproduce that material in responses to users.

ANI’s case is also being closely watched as an early Indian legal challenge involving ChatGPT and the use of news content by generative AI.

OpenAI has challenged the maintainability of ANI’s suit in India. The company has argued that its AI models are not trained in India and that its relevant servers are located in the United States.

The company has also maintained that its machine-learning process is transformative and that facts contained in news reports are not themselves protected by copyright.

OpenAI further argued that publishers that do not want their content crawled can request to be placed on a blocklist.

These arguments put the dispute at the intersection of copyright law, AI training practices and the territorial reach of Indian courts.

The single judge had rejected ANI’s application seeking an interim injunction against OpenAI.

The court held that OpenAI’s storage of ANI’s original works did not, by itself, amount to copyright infringement under the Copyright Act, 1957.

On the research-related defence raised in the case, the court had also observed that training the large language models underlying ChatGPT involves machine learning of stored literary works through processes including screening and organising the material.

ANI’s latest appeal seeks to challenge that interim-stage decision, bringing the copyright dispute before a division bench of the Delhi High Court.

The September 14 hearing could therefore provide an important next step in the wider debate over how Indian copyright law applies to generative AI systems and the news material used by them.