Skip to main content

Indian Court Rules for OpenAI: Training AI on News Is Fair Use

A Delhi High Court judgment has settled a key question in India: does training AI on news articles require paying for content? The answer, delivered on July 24, is a clear no. Judge Amit Bansal ruled that OpenAI's use of content from Asian News International (ANI) to train its AI model falls under fair use and does not constitute copyright infringement.

The court gave two main reasons. First, OpenAI's actions are protected by the fair use exception in India's Copyright Act. Second, ANI failed to provide evidence that ChatGPT's responses directly copied ANI's copyrighted material. These points dismantled the foundation of ANI's lawsuit.

More significantly, the judge rejected ANI's request for a temporary injunction. He emphasized that such an order would not help AI development but would instead severely hinder India's emerging large language model projects. It would also harm the public interest, affecting users who freely access ChatGPT in India. The court weighed protecting one content provider against hurting the entire industry and the public, and chose the latter.

This ruling is not an isolated event. Media outlets like The New York Times have filed similar lawsuits against OpenAI in the United States, with cases still pending. As these disputes unfold in different jurisdictions, India's decision has opened a crack in the argument that training AI on news automatically equals infringement.

Key Points

  • Fair Use Victory: The Delhi High Court ruled that OpenAI's use of news content for AI training is fair use under Indian copyright law.
  • No Evidence of Copying: ANI failed to show that ChatGPT directly copied its copyrighted content.
  • Injunction Denied: The court rejected a temporary injunction, warning it would harm India's AI industry and public access.
  • Global Context: Similar lawsuits against OpenAI are ongoing in the US, making this a significant precedent.