Delhi High Court rules 'private use' not AI safe harbour in copyright case

This is a judgment by the High Court of Delhi (Justice Amit Bansal) in a copyright suit filed by ANI Media Pvt. Ltd. against Open AI, addressing whether using copyrighted news data to train AI models constitutes infringement and whether it qualifies as 'fair use' under Indian law. It matters to citizens as it establishes legal boundaries for AI training on copyrighted content in India.

Who it concerns: ANI Media Pvt. Ltd. (Plaintiff), Open AI (Defendant), and various intervenors including the Federation of Indian Publishers, Broadband India Forum, and Flux AI Labs.

What the document says

• The court examined whether Open AI's use of ANI's copyrighted data for training its LLM (ChatGPT) amounts to copyright infringement under the Copyright Act, 1957.
• The court analyzed whether the use qualifies as 'fair use' under Section 52(1)(a) of the Copyright Act, specifically considering 'private or personal use, including research'.
• The court assessed whether ChatGPT's responses constitute a substantial reproduction of ANI's copyrighted literary works.
• The court determined its territorial jurisdiction to entertain the suit despite the defendant's servers being located in the United States.

Also in the document

• The judgment was reserved on 27th March, 2026, and pronounced on 24th July, 2026.
• The case involves two specific claims: a 'training claim' (copying/storage for training) and a 'reproduction or output claim' (reproduction in AI responses).
• The court noted that India is considered a forerunner in AI, citing the launch of Sarvam AI's domestic LLM in February 2026.
• The judgment discusses the legislative history and interpretation of Section 52(1)(a) of the Copyright Act, 1957, including the 'purpose test' and 'fairness test'.