The Delhi High Court has ruled, at the interim stage, that OpenAI’s use of news agency ANI’s copyrighted content to train ChatGPT qualifies as “fair dealing” under India’s Copyright Act, 1957, marking the country’s first major judicial decision on the legality of using copyrighted material to train generative AI models. Justice Amit Bansal declined to grant an interim injunction sought by ANI, holding that OpenAI had made a prima facie case that storing and using ANI’s content for training large language models falls within the statutory exception for “private or personal use, including research” under Section 52(1)(a) of the Copyright Act.
The ruling is a significant milestone for India’s AI ecosystem, providing preliminary legal support for AI developers training models on publicly available copyrighted works. However, the court emphasized that the decision relates only to interim relief, meaning the underlying copyright lawsuit will continue, and the final legal position will be determined after a full trial.
Delhi High Court Refuses ANI’s Request for Interim Injunction
ANI filed a copyright infringement lawsuit against OpenAI in 2024, alleging that the company had:
- Used ANI’s copyrighted news reports without permission to train ChatGPT.
- Stored its copyrighted works in OpenAI’s training datasets.
- Generated inaccurate responses falsely attributed to ANI, harming its reputation.
ANI sought an interim injunction to stop OpenAI from using its content while the case proceeds. The High Court rejected that request, finding that ANI had not established a prima facie case warranting such relief.
Case Snapshot
| Item | Details |
|---|---|
| Court | Delhi High Court |
| Judge | Justice Amit Bansal |
| Plaintiff | ANI Media Pvt. Ltd. |
| Defendant | OpenAI |
| Main Issue | Use of copyrighted news content to train ChatGPT |
| Current Status | Interim injunction denied; main case continues |
Court Holds AI Training Is Prima Facie ‘Fair Dealing’
The court concluded that OpenAI’s storage of ANI’s literary works solely for training its large language models is, at least on a prima facie basis, protected under Section 52(1)(a) of the Copyright Act.
Justice Amit Bansal observed that:
- Training AI models constitutes research within the meaning of the Copyright Act.
- The use is transformative because the content is used to improve an AI model rather than substitute for the original works.
- Such use promotes technological advancement and public interest.
- Therefore, OpenAI’s conduct appears to fall within India’s fair dealing exception.
The judgment marks the first substantial interpretation by an Indian court of how copyright law applies to generative AI training.
ANI Failed to Show Copyright Reproduction
Another important finding concerned ChatGPT’s outputs.
The court found that ANI had not demonstrated that:
- ChatGPT memorized ANI’s articles.
- The AI reproduced substantial portions of ANI’s original works.
- Users received outputs substantially similar to ANI’s copyrighted content.
Without evidence of direct reproduction, the court found no prima facie basis for copyright infringement arising from ChatGPT’s responses, including those generated using Retrieval-Augmented Generation (RAG).
Court’s Key Findings
| Issue | Court’s Prima Facie View |
|---|---|
| AI training on copyrighted works | Protected as fair dealing for research |
| Storage of ANI content | Not copyright infringement |
| ChatGPT outputs | Not substantially similar to ANI’s reports |
| Interim injunction | Refused |
Why the Ruling Matters
The ruling has been widely covered for its broader implications, with related coverage confirming the Delhi High Court’s finding that OpenAI’s use of ANI content is not copyright infringement.
The decision is significant because it addresses one of the biggest legal questions facing the AI industry: whether copyrighted material can be used to train foundation models without obtaining licenses.
The ruling could:
- Provide greater legal certainty for AI developers operating in India.
- Influence future copyright litigation involving generative AI.
- Encourage AI research and innovation under existing copyright exceptions.
- Shape how courts interpret “fair dealing” in the context of machine learning.
At the same time, the judgment does not establish a final legal precedent, as the court’s observations were made while deciding whether interim relief should be granted.
Case Is Far From Over
Although OpenAI secured an important interim victory, the litigation continues.
The Delhi High Court will still examine:
- Whether OpenAI’s conduct ultimately infringes ANI’s copyright.
- The scope of India’s fair dealing provisions for AI training.
- Broader issues relating to AI, copyright, licensing, and digital publishing.
The court also confirmed that it has jurisdiction to hear the dispute, allowing the case to proceed on its merits.
Global Context
AI copyright disputes remain active elsewhere too, as seen when the New York Times accused OpenAI of hiding evidence in its ChatGPT copyright trial.
The ANI lawsuit is part of a broader wave of copyright litigation against AI companies worldwide.
OpenAI is facing similar legal challenges from:
- News organizations.
- Book publishers.
- Authors.
- Media companies.
Courts in the United States, Canada, and Europe are also considering whether AI companies can train large language models on copyrighted material without permission, making the Delhi High Court’s ruling an important contribution to the evolving global legal debate.
Looking Ahead
The Delhi High Court’s interim ruling represents a landmark moment for India’s artificial intelligence and copyright landscape. By holding that OpenAI’s use of ANI’s copyrighted news content for training ChatGPT is, prima facie, protected under the Copyright Act’s fair dealing provisions for research, the court has offered early judicial support for AI model training using publicly available material. It also found no evidence at this stage that ChatGPT reproduced ANI’s articles in a manner amounting to copyright infringement.
Looking ahead, the final outcome of the case will be closely watched by AI developers, publishers, media organizations, and policymakers worldwide. A final judgment could define the boundaries of AI training under Indian copyright law and influence future legislation, licensing practices, and judicial decisions as countries continue to grapple with balancing technological innovation and the rights of copyright owners.
Frequently Asked Questions
What did the Delhi High Court rule on OpenAI and ANI?
The Delhi High Court ruled, at the interim stage, that OpenAI’s use of news agency ANI’s copyrighted content to train ChatGPT qualifies as fair dealing under India’s Copyright Act, 1957, and declined to grant ANI an interim injunction.
Why did the court allow OpenAI’s use of ANI’s content?
Justice Amit Bansal held that OpenAI made a prima facie case that storing and using ANI’s content for training large language models falls within the statutory exception for private or personal use, including research, under Section 52(1)(a) of the Copyright Act.
Is the legal case over now?
No, the case is far from over. This was an interim ruling denying ANI’s request for an injunction, and the underlying dispute remains to be fully decided.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
