The Delhi High Court dismissed ANI's bid for an immediate stay, allowing OpenAI to continue training large language models on Indian content without a licence. The ruling spotlights the clash between Indian copyright law and rapid AI development.

  • ANI's request for an immediate stay was denied by the court.
  • OpenAI can keep training on Indian-origin content without a licence.
  • The case is slated for a further hearing on December 8.

Case Background

In 2024, news agency ANI sued OpenAI Inc. and its subsidiary, alleging unlawful use of its copyrighted articles, images, and video footage to train large language models (LLMs). It was the first high‑profile AI‑training dispute in Indian courts.

A single judge in July 2026 held that using ANI's works for LLM training fell under “private” use, including research, and therefore qualified as fair use, meaning no infringement occurred. This decision opened the door for AI developers to scrape Indian content without licensing agreements.

Read Also:Wedding Calls Off in Punjab's Firozpur After Bride Rejects Groom Upon Seeing Him

Recent Delhi High Court Action

On 15 September 2026, a division bench comprising Justices Avneesh Jhingan and Manmeet Pritam Singh Arora briefly heard ANI’s appeal. The judges verbally noted that they would not pass any ex‑parte order and asked OpenAI to file a response. The matter is scheduled for a full hearing on 8 December.

Historical Background

India’s Copyright Act, 1957, provides a narrow “fair dealing” exception that does not mirror the broader U.S. fair‑use doctrine. Consequently, the legal framework offers limited clarity for AI developers seeking to use copyrighted material at scale, prompting media houses like ANI to protect their statutory rights in court.

Why This Matters

BozokMedia analysis shows that the outcome will influence not only Indian media economics but also global AI policy debates, as jurisdictions worldwide grapple with similar copyright‑training conflicts.

"Unlicensed large‑scale training on copyrighted works could erode creators' economic rights," says digital‑rights scholar Dr. Anita Gupta.
Did You Know?: OpenAI told the court it had already "blocklisted" ANI’s domain in November 2024, removing it from future training datasets.

Frequently Asked Questions

Q1: How much licence fee did ANI propose to OpenAI?
A: USD 7.5 million.

Q2: When is the next hearing scheduled?
A: 8 December 2026.