By Aditya Kalra

NEW DELHI (Reuters) - A group of India's top Bollywood music labels, from T-Series to Saregama and Sony ( SONY ), is seeking to join a copyright lawsuit against OpenAI in New Delhi, highlighting worries about improper use of recordings to train AI models, legal documents show.

Microsoft ( MSFT )-backed OpenAI's legal challenges are mounting globally and in India, its second biggest market by users. But the company says it follows fair-use principles in employing publicly available data to build its AI models.

On Thursday, the Indian Music Industry (IMI) group, T-Series and Saregama India asked a New Delhi court to hear concerns about "unauthorised use of sound recordings" in training AI models that breaches their copyright.

The companies' contentions in the lawsuit "are crucial for the entire music industry in India, and even worldwide," they said in their filing, which is not public but was reviewed by Reuters.

OpenAI and the music labels did not respond to requests for comments on Friday.

The music labels want to join a lawsuit launched last year by Indian news agency ANI that accused OpenAI's ChatGPT application of using its content without permission to train AI models.

Since then, book publishers and media groups, some backed by billionaires Mukesh Ambani and Gautam Adani, have banded together to oppose the company in the New Delhi court.

Bollywood and Hindi pop music are big business in India.

T-Series is one of India's largest music record labels which releases about 2,000 sound records or songs annually, while Saregama, more than 100 years old, owns a repertoire of famed Indian singers such as Mohammed Rafi and Lata Mangeshkar.

On its website, the IMI group says it also represents global names such as Sony Music and Warner Music ( WMG ).

In India, the music labels are "concerned OpenAI and other AI systems can extract lyrics, music compositions and sound recordings from the internet," said an industry source who spoke on condition of anonymity as the matter is in court.

The Indian companies' latest action comes after Germany's GEMA, which represents composers, lyricists and publishers, said in November it had sued OpenAI for ChatGPT's alleged unlicensed reproduction of song lyrics with which "the system has obviously been trained".

OpenAI, which is grappling with new challenges from Chinese startup DeepSeek's breakthrough in cheap AI computing, opposed the ANI lawsuit on the grounds that Indian courts lack jurisdiction, as the company is U.S.-based, with servers abroad.

The next hearing in the lawsuit, which is seen as shaping the future of how AI models use copyright content in India, is set for Feb. 21.

OpenAI chief Sam Altman visited India last week, meeting the infotech minister, and discussing the country's plan to pursue low-cost AI.