ARTICLE
2 September 2026

Copyrights – Use Of Publicly Available Content Of A News Agency For Training An AI Model Prima Facie Falls Under ‘fair Dealing’

LS
Lakshmikumaran & Sridharan

Contributor

Lakshmikumaran & Sridharan (LKS) is a premier full-service Indian law firm specializing in areas such as corporate & M&A/PE, dispute resolution, taxation and intellectual property. The firm, through its 14 offices across India works closely on litigation and commercial law matters, advising and representing clients both in India and abroad.
The Delhi High Court has declined interim relief in a copyright infringement suit brought by a news agency against an AI developer, finding that the use of publicly available content for training AI models may constitute fair dealing under Indian copyright law. The Court examined whether AI-generated outputs substantially reproduced copyrighted works and whether such use serves broader public interests in innovation and access to information.
India Intellectual Property
Lakshmikumaran & Sridharan are most popular:
  • within Insurance, Transport and Environment topic(s)

In an intersection of copyright law and artificial intelligence, the Delhi High Court has declined to grant interim relief sought by a news agency in a copyright infringement suit against an AI developer and service provider. The Court was of the prima facie view that the use of publication house’s publicly available content for training the AI model fell within the ‘fair dealing’ protection under Section 52(1)(a) of the Copyright Act, 1957.

The Court in ANI Media Pvt. Ltd. v. OpenAI OpCo LLC noted that the publication house/petitioner failed to establish that AI model’s outputs substantially reproduced its copyrighted works or that the AI models memorised and regurgitated the content. Dismissing the interim injunction application, the High Court noted the following:

  • It had territorial jurisdiction to hear the dispute despite the respondent’s (AI service provider’s) servers being located outside India. It was noted that the copyrighted work was accessed from India.
  • The news site could not demonstrate any substantial similarity between its news articles and respondent’s AI-generated responses sufficient to constitute copyright infringement.
  • Copyright does not subsist in facts or news events themselves, but only in their original expression.
  • Temporary or permanent storage of copyrighted works for AI training amounts to reproduction under Section 14 of the Copyright Act, but such use may be protected by the fair dealing exception under Section 52.
  • Training large language models constitutes ‘private or personal use, including research’ and satisfies the requirement of the statutory exception under Section 52(1)(a).
  • Commercial use automatically does not disqualify a party from claiming fair dealing protection under Section 52(1)(a).
  • Under the Explanation to Section 52(1)(a), the limitation of not being an infringing copy is not applicable in respect of the works stored in electronic mode.
  • AI service provider’s use was limited to training its models and it did not function as a substitute for the articles published by the news site of the petitioner. It served broader public interests such as research, innovation, education and access to information.
  • Use of news agency’s works by the AI model will not affect its market share or cause actual or potential damage to the news agency.
  • Public interest – An adverse order could impede AI innovation and affect millions of users in India.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.

[View Source]

Mondaq uses cookies on this website. By using our website you agree to our use of cookies as set out in our Privacy Policy.

Learn More