OpenAI faces mounting lawsuits in copyright row with Indian news outlets

OpenAI’s legal challenges in India could escalate as several major news outlets including Hindustan Times, Indian Express, and The Hindu reportedly seek to join an ongoing copyright lawsuit against the Microsoft-backed AI startup. 

The lawsuit, which accuses OpenAI of using their news content without permission to train generative AI model ChatGPT, was filed by Indian news agency ANI last November. 

Digital News Publishers Association (DNPA), which represents 20 media outlets  including Hindustan Times, Indian Express, Network18, The Hindu, India Today, Zee News and Dainik Jagran  has filed a 135-page appeal in Delhi High Court alleging that OpenAI’s actions pose a threat to “valuable copyright data” of news publishers in India, according to a Reuters report. 

OpenAI has refuted the copyright infringement allegations, maintaining that its AI models are trained on publicly available data, which it argues qualifies as fair use under the US law. This is because the content generated by ChatGPT does not replicate the original work verbatim. 

Last week, OpenAI informed the Delhi High Court that it doesn’t fall under the jurisdiction of Indian courts as the company has no office or legal presence in India. However, in April 2024, the AI startup appointed Pragya Mishra, former WhatsApp India communications manager, as its public policy and partnership head. 

In response to ANI’s appeal to delete all news data generated by the news agency, OpenAI told the Delhi High Court that deleting any data will be in violation of US laws that require companies to preserve all data during litigation. OpenAI is currently facing copyright infringement lawsuits by several US news publishers including New York Times, News York Daily News and Centre for Investigative Reporting (CIR). 

OpenAI has been accused in these lawsuits of using millions of copyrighted articles without permission or payment to the news outlets to create competing products, which not only infringes on their intellectual property but also threatens their news business.  

Despite these legal challenges, OpenAI has managed to strike content sharing deals with many other global news publishers including Associated Press, Vox Media, Axel Springer and News Corp, and Financial Times, which gives the AI startup the license to use select news content for training AI models and answering user queries. 

Indian news publishers in their filing have highlighted OpenAI’s partnerships with global news outlets, indicating an interest in pursuing similar agreements that involve a licensing fee. OpenAI’s payment to news outlets for their copyrighted content varies based on the value of the content and size of the outlet. For example, NewsCorp, which owns publications such as Wall Street Journal, Dow Jones & Company and The Times (UK), is reportedly set to make $250 million over a period of five years, while the deal with Financial Times is valued at $5 million to $10 million. 

Multiple AI firms offering generative AI products are facing legal challenges over their data collection practices. For instance, US based startup Perplexity AI, which offers a generative AI powered search app, was sued last October by The Wall Street Journal, New York Post, Financial News, and Barron’s for allegedly using their copyrighted news articles to train its AI models. 

Microsoft and OpenAI are also facing legal action over copyright infringement from the Authors’ Guild and several prominent writers including David Baldacci, John Grisham, George R.R. Martin, and Michael Connelly. 

Currently valued at $157 billion, OpenAI is reportedly planning to restructure itself and transition into a for-profit benefit corporation. This new entity will operate independently under CEO Sam Altman and will not be controlled by OpenAI’s non-profit board. Last week, SoftBank and Oracle announced plans to invest $500 billion under the Stargate Project over the next four years to build new AI infrastructure for OpenAI in the US. 

Image credit: Pexels