An Indian court has ruled that OpenAI’s use of content belonging to news agency Asian News International for training ChatGPT did not constitute copyright infringement, handing the artificial-intelligence company a significant victory in one of India’s most closely watched AI-related legal disputes.
The court delivered its decision on Friday in a lawsuit filed by ANI, which had accused OpenAI of using its copyrighted news content without authorisation to train the large language models supporting ChatGPT.
ANI had also raised concerns about ChatGPT allegedly generating fabricated responses and attributing statements or interviews to the news agency that it had not published. The agency had sought damages and the removal of its content from OpenAI’s training data.
OpenAI rejected the allegations and argued that training its artificial-intelligence models on publicly available information was legally permissible. The company also maintained that its systems do not store or reproduce training material in the same form as conventional databases. It had previously stated that ANI’s website was blocked from future web crawling after the agency exercised an opt-out option.
The dispute began in 2024 and subsequently attracted participation from Indian and international publishing organisations concerned about how generative-AI companies acquire and use copyrighted material. It required the court to consider whether using protected content for AI training amounted to infringement under Indian copyright law.
Friday’s ruling could influence how Indian courts approach similar disputes involving AI developers, news organisations, publishers and other content owners. However, the detailed reasoning and precise scope of the decision will need to be examined through the court’s written judgment. The ruling concerns OpenAI’s use of ANI’s material in this specific case and should not automatically be interpreted as unrestricted permission for every AI company to use copyrighted content.
The decision arrives as courts in several countries examine whether training generative-AI systems on copyrighted books, news reports, images and other material requires permission or licensing. Technology companies generally argue that model training transforms source material, while publishers and creators maintain that unauthorised use undermines their rights and commercial licensing models.
