by Shradhanjali Sarma & Shatakshi Shekhar
New Delhi: In a landmark order that could influence the future of artificial intelligence regulation in India, the Delhi High Court has refused to grant interim relief to news agency ANI in its copyright infringement suit against OpenAI. The ruling, delivered on July 24, is being seen as India’s first significant judicial pronouncement on generative AI and copyright. However, the court made it clear that it has not ruled on whether the use of copyrighted material for training AI models is lawful.
What is the dispute?
ANI alleged that OpenAI used its copyrighted news reports to train ChatGPT without authorisation and that the chatbot was capable of reproducing portions of its content, amounting to copyright infringement. The agency sought an interim injunction restraining OpenAI from using its copyrighted material while the case is pending.
What did the Delhi High Court rule?
Rejecting ANI’s plea for interim relief, the Delhi High Court held that the news agency had not, at this preliminary stage, produced sufficient evidence to establish that ChatGPT had “memorised” or reproduced its reports in a substantially similar form.
The court observed that the storage of copyrighted material for AI training could potentially fall within the Copyright Act’s exception for “private or personal use, including research.” It further noted that the chatbot responses presented before it were not substantially similar to ANI’s original reports.
The court also rejected OpenAI’s argument that Indian courts lacked jurisdiction because its servers are located in the United States, allowing the copyright suit to continue in India. It added that suspending ChatGPT’s services at the interim stage could adversely affect the public given the platform’s widespread use.
What the ruling means — and what it doesn’t
Legal experts have cautioned against interpreting the order as a blanket approval for AI companies to use copyrighted material for training their models.
The High Court did not hold that AI developers are legally entitled to train AI systems using copyrighted content. It merely concluded that ANI had failed to present enough evidence at this stage to justify the grant of an interim injunction.
The larger legal questions relating to AI training, copyright infringement, fair use and the scope of protection under India’s Copyright Act remain undecided and will be examined during the trial.
Why the case matters
The ANI-OpenAI dispute is India’s first major courtroom battle over generative AI and copyright. The order suggests that future AI copyright litigation may hinge on concrete evidence demonstrating whether an AI system can reproduce copyrighted works in a substantially similar form, rather than solely on the broader issue of whether AI training itself constitutes copyright infringement.
Legal commentators have also drawn comparisons with Germany’s GEMA v. OpenAI case, where the court found evidence that ChatGPT could reproduce copyrighted song lyrics almost verbatim. According to experts, the differing outcomes reflect differences in the evidence presented rather than conflicting legal principles.
With interim relief denied, the case will now proceed to trial, where the Delhi High Court is expected to examine the legal boundaries governing AI model training, copyright protection and the use of copyrighted works in generative AI systems.
Notably, the legal analysis has been contributed by Shradhanjali Sarma, Founder Partner at Sakura Law Chambers, and Shatakshi Shekhar, Lead – Product Policy & Government Affairs at Sakura Law Chambers.
