The Delhi High Court has declined, for now, to stop OpenAI from using ANI material to train ChatGPT, marking an early and closely watched legal development in India's AI copyright debate. The court's interim view was that the use in question falls under fair dealing for research, based on the material placed before it at this stage.
According to the ruling, the court did not find proof that ChatGPT had reproduced ANI's news reports in a way that would justify immediate restraint. That point appears central to the decision, because the dispute is not only about whether copyrighted news content was used in training, but also whether the AI system later generated outputs that copied protected reporting.
Justice Amit Bansal's order is being seen as one of India's first meaningful judicial responses to the fast-growing clash between publishers and AI companies. The case highlights a broader issue facing courts worldwide: whether training large language models on copyrighted material should be treated as infringement or as a legally permitted research-related use.
While the order does not end the dispute, it gives OpenAI room to continue for the moment and signals how Indian courts may begin approaching AI training claims. The larger questions around copyright, news content, and the boundaries of fair dealing in the age of generative AI remain open and are likely to be tested further as the case proceeds.
