The Delhi High Court has ruled that OpenAI’s use of Asian News International (ANI) content to train ChatGPT does not amount to copyright infringement, delivering the first major judicial decision in India on the legality of using copyrighted material to train artificial intelligence models. Justice Amit Bansal held that ANI failed to demonstrate that ChatGPT reproduced or memorized its copyrighted news reports in a manner that violated the Copyright Act, while also finding that the use of publicly available material for AI training qualified as “fair dealing for research” under Indian copyright law.
The judgment is a significant milestone for India’s AI ecosystem and could shape future litigation involving generative AI developers, publishers, and copyright owners. The court also rejected ANI’s request for an injunction that would have prevented OpenAI from using its publicly available news content to train its models, although separate issues relating to allegedly false attribution of news reports remain distinct from the copyright ruling.
Delhi High Court Rules in OpenAI’s Favour
The case stemmed from a lawsuit filed by ANI alleging that OpenAI had used its copyrighted news articles without permission to train ChatGPT.
Justice Amit Bansal ruled that:
- AI training on ANI’s publicly available content did not constitute copyright infringement.
- ANI failed to establish that ChatGPT reproduced protected expression from its articles.
- The use qualified as “fair dealing for research” under India’s Copyright Act.
- An injunction preventing OpenAI from using ANI’s content was therefore not warranted.
Case Overview
| Item | Details |
|---|---|
| Plaintiff | Asian News International (ANI) |
| Defendant | OpenAI |
| Court | Delhi High Court |
| Judge | Justice Amit Bansal |
| Core Issue | Copyright infringement through AI training |
| Outcome | Court ruled in favour of OpenAI |
Court’s Reasoning
The court concluded that copyright law protects the expression contained in a work, not the underlying facts or information.
According to the judgment:
- ANI did not prove that ChatGPT stored or reproduced substantial portions of its articles.
- Transforming publicly available text into model parameters for training was not shown to infringe copyright in this case.
- The training process fell within the scope of research-related fair dealing under Indian law.
Background of the Dispute
ANI filed the lawsuit in November 2024, alleging that OpenAI:
- Used its news reports without obtaining a licence.
- Trained ChatGPT on copyrighted content.
- Generated fabricated news reports attributed to ANI.
- Benefited commercially from copyrighted works without authorization.
OpenAI denied the allegations and argued that:
- Its models were trained using publicly available information.
- The training process was lawful.
- It did not reproduce copyrighted articles in a manner that infringed copyright.
Main Arguments
| ANI | OpenAI |
|---|---|
| AI training used copyrighted articles without permission | Training relied on publicly available information |
| Commercial use required licensing | Training process was legally protected |
| ChatGPT generated false ANI-attributed reports | Copyright was not infringed through model training |
Significance for India’s AI Industry
The judgment is expected to have implications beyond the OpenAI-ANI dispute.
It may influence:
- Future copyright litigation involving AI developers.
- Licensing negotiations between publishers and AI companies.
- India’s approach to AI regulation.
- Development of generative AI products within the country.
The decision could also be referenced in ongoing AI copyright disputes in other jurisdictions, although each country’s copyright framework differs.
Why It Matters
This is the first major Indian court ruling to address whether training a generative AI model on copyrighted news content amounts to copyright infringement. By finding that OpenAI’s use of ANI’s publicly available material qualified as fair dealing for research and that ANI had not shown unlawful reproduction of its protected expression, the Delhi High Court has established an important legal precedent for AI developers operating in India.
The decision is also likely to influence how publishers and technology companies approach licensing negotiations, AI training practices, and future litigation. While the ruling strengthens OpenAI’s position in India, it does not necessarily resolve every legal issue surrounding AI and copyrighted content, and additional disputes involving different facts or legal claims may still emerge as generative AI adoption continues to expand.
Looking Ahead
The Delhi High Court’s judgment is expected to become a key reference point in India’s evolving legal framework for artificial intelligence and copyright. Publishers, AI developers, and policymakers will closely examine the reasoning as they navigate questions around text and data mining, licensing arrangements, and the balance between innovation and intellectual property protection.
Looking ahead, the case may encourage further legislative and judicial clarification on AI training under Indian copyright law. As generative AI becomes increasingly integrated into business and consumer applications, courts and lawmakers are likely to face additional questions about fair dealing, model training, attribution, and the commercial use of copyrighted works.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
