OpenAI Denies Using Indian Media Content To Train ChatGPT

In the latest twist in the copyright infringement case against OpenAI, the AI giant has reportedly informed the Delhi High Court (HC) that it does not use the content of Indian media groups to train its AI chatbot ChatGPT.

As per a court filing reviewed by Reuters, OpenAI has urged the HC to dismiss infringement claims filed by the likes of Mukesh Ambani-owned Network18, Gautam Adani-owned NDTV and others. 

OpenAI has also told the HC that it is under no obligation to form partnerships with media outlets to use publicly available content. Notably, the filing is part of a broader lawsuit initially filed by Indian news agency ANI last year. 

ANI claimed that OpenAI was using its published content without permission to train its chatbot. 

Previously, the AI giant told the HC that it cannot delete the training data as it is currently defending a litigation in the US, adding that US laws require the company to preserve the data while the hearings are pending.

However, since ANI filed the case last year, many industry bodies of book publishers (via Federation of Indian Publishers) and news channels (under the banner of Digital news Publishers Association) have also sued the AI giant for allegedly using their data to train ChatGPT in violation of copyright laws. 

In its latest filing dated February 11, OpenAI denied using any content from the media groups to train its models.

In stark contrast with India, OpenAI has reportedly struck content display deals with some news outlets globally. Local media groups argue that no similar agreements have been made by OpenAI in India. 

For instance, in August 2024, OpenAI signed a deal with the US-based publisher Condé Nast, which allowed the AI giant to use content from The New Yorker, Vogue, GQ, Vanity Fair and Bon Appétit etc. Under the said deal, OpenAI now has permission to display content from Condé Nast’s media properties in its products, including ChatGPT and SearchGPT. 

In its latest filing with Delhi HC, OpenAI has argued that its other partnerships abroad are not just “licensing arrangements for the purpose of training” its AI models. The company has also said, in the filing, that the use of content, which is publicly available, is permissible under Indian copyright law.

The court hearing in the ANI’s lawsuit is scheduled to take place next week.

Earlier, OpenAI told the Delhi HC that Indian courts do not have the jurisdiction to hear ANI’s copyright infringement case as the company does not have a presence in the country. 

The latest development comes barely a week after OpenAI chief Sam Altman visited India, where he quipped that “India should be one of the leaders of the AI revolution”.

The post OpenAI Denies Using Indian Media Content To Train ChatGPT appeared first on Inc42 Media.