Delhi High Court Rules Training on Copyrighted News Is 'Fair Dealing' and RAG Outputs Don't Infringe

Bar and Bench (corroborated by Business Standard, LiveLaw)·high signal

On July 24 Justice Amit Bansal denied ANI Media interim relief against OpenAI, holding that training LLMs on ANI's copyrighted content falls within Section 52(1)(a)(i) "private or personal use, including research" and constitutes fair dealing. The court separately found that ChatGPT outputs produced via retrieval-augmented generation do not infringe under Section 51 because they were not substantially similar to ANI's original literary works. ANI was the first Indian media house to sue OpenAI, and the RAG holding is the part builders should note — it treats retrieval-grounded generation as a distinct question from training, in the world's largest ChatGPT user market.