The Delhi High Court has reserved its order on a petition filed by news agency Asian News International (ANI), which seeks to restrain OpenAI from using its copyrighted content to train ChatGPT. The case has emerged as a landmark legal battle, marking India’s first major judicial examination of how existing copyright law applies to artificial intelligence systems.
The matter was heard by Justice Amit Bansal over the course of 32 hearings, spanning from November 2024 to March 2026. What began as a dispute between a news agency and a technology company has evolved into a broader legal confrontation with far-reaching implications for the media industry, digital publishing, and the development of generative AI in India.
ANI has alleged that OpenAI used its copyrighted news articles without permission or licensing, arguing that the content was scraped and incorporated into the training data of ChatGPT. It further claimed that the system is capable of reproducing portions of its reports in responses, including material that was behind paywalls and accessible only to subscribers. According to ANI, such usage does not fall under the “fair dealing” exception provided under Indian copyright law, especially given the commercial nature of AI products.
The agency also contended that copyright infringement occurs at the initial stage of data ingestion itself. It argued that processes such as scraping, tokenisation, and storage of content involve unauthorised use of protected material, regardless of whether the final output reproduces it verbatim.
OpenAI, in its defence, challenged both the jurisdiction of Indian courts and the substance of the allegations. It argued that since the company and its servers are not based in India, the case may not be maintainable. On the merits, OpenAI stated that its models are trained on publicly available data and that the training process focuses on extracting statistical patterns rather than copying expressive content.
The company also maintained that any storage of data during training is temporary and incidental, and therefore does not constitute infringement. It further informed the court that it has already blocked ANI’s domain from being included in future training datasets. Additionally, OpenAI argued that ANI had not demonstrated clear instances of verbatim reproduction of its content in ChatGPT’s outputs.
The case has attracted multiple intervenors representing different sectors. Organisations such as the Digital News Publishers Association, the Indian Music Industry, and the Federation of Indian Publishers supported ANI’s position, warning that large-scale unauthorised use of content could undermine the economic viability of creative industries. They argued that AI training involves systematic copying and that AI-generated outputs could act as substitutes for original works, thereby causing market harm.
On the other side, entities including the Broadband India Forum, Flux Labs AI, and IGAP Project LLP supported OpenAI, advocating for a legal interpretation that accommodates technological innovation. They argued that large language models do not store or reproduce content in a traditional sense and that transforming or summarising publicly available information should not be treated as copyright infringement.
At the core of the dispute are critical legal questions that could define the future of AI regulation in India. One key issue is whether intermediate copies created during AI training can be considered “transient or incidental storage” under Indian law, or whether they amount to unlawful reproduction. Another major point of contention is the scope of the “fair dealing” exception under Section 52 of the Copyright Act, particularly whether it can extend to commercial AI systems.
Even court-appointed amici curiae presented differing views, reflecting the complexity of the issue. While one opinion suggested that OpenAI’s practices constitute infringement, another indicated that certain aspects of AI training may fall within permissible legal boundaries.
The High Court’s forthcoming decision on interim relief is expected to offer the first clear judicial direction on how India may approach the intersection of copyright law and artificial intelligence. Its impact is likely to extend beyond this case, shaping how AI systems are developed, trained, and regulated, while also influencing the balance between innovation and the protection of intellectual property rights in the digital age.
