India's biggest news organizations are seeking to join a lawsuit against OpenAI, the US startup behind ChatGPT, for alleged unauthorised use of their content. The news organizations include some of India's oldest publications like The Indian Express, The Hindu, The India Today group, billionaire Gautam Adani-owned NDTV, and over a dozen others. OpenAI denies the allegations and told the BBC that it uses "publicly available data" that are in line with "widely accepted legal precedents". On Wednesday, OpenAI CEO Sam Altman was in Delhi to discuss India's plan for a low-cost AI ecosystem with IT Minister Ashwini Vaishnaw. He said India "should be one of the leaders of the AI revolution" and said earlier comments from 2023, when he said Indian firms would struggle to compete, had been taken out of context. "India is an incredibly important market for AI in general and for OpenAI in particular," local media quoted him as saying at the event. The legal case filed against OpenAI in November by Asian News International (ANI), India's largest news agency, is the first of its kind in India. ANI accuses ChatGPT of using its copyrighted material illegally — which OpenAI denies — and is seeking damages of 20m rupees ($230,000; £185,000). The case holds significance for ChatGPT given its plans to expand in the country. According to a survey, India already has the largest user base of ChatGPT. Chatbots like ChatGPT are trained on massive datasets collected by crawling through the internet. The content produced by nearly 450 news channels and 17,000 newspapers in India holds huge potential for this. There is, however, no clarity on what material ChatGPT can legally collect and use for this purpose. OpenAI is facing at least a dozen lawsuits across the world filed by publishers, artists and news organizations, who have all accused ChatGPT of using their content without permission. The most prominent of them was filed by The New York Times in December 2023, in which the newspaper demanded "billions of dollars" in damages from OpenAI and Microsoft, its backer. "A decision by any court would also hold some persuasive value for other similar cases around the world," says Vibhav Mithal, a lawyer specialising in artificial intelligence at the Indian law firm Anand and Anand. Mithal said the verdict in the lawsuit filed by ANI could "define how these AI models will operate in the future" and "what copyrighted news content can be used to train AI generative models [like ChatGPT]". A court ruling in ANI's favor could spark further legal cases as well as opening the possibility of AI companies entering into license sharing agreements with content creators, which some companies have already started doing. "But a ruling in OpenAI's favour will lead to more freedom to use copyrighted protected data to train AI models," he said. ANI provides news to its paying subscribers and owns exclusive copyright over a large archive of text, images and videos. In its suit filed in the Delhi High Court, ANI says that OpenAI used its content to train ChatGPT without permission. ANI has argued that this led to the chatbot getting better and has profited OpenAI. The news agency said that before filing the suit, it had told OpenAI its content was being used unlawfully and offered to grant the company a license to use its data. ANI says OpenAI declined the offer and put the news agency on an internal blocklist so that its data is no longer collected. It also asked ANI to disable certain web crawlers to ensure that its content was not picked up by ChatGPT. The news agency says that despite these measures, ChatGPT picks up its content from websites of its subscribers. This has enriched OpenAI "unjustly", it says. ANI also says in its suit that the chatbot produces its content verbatim for certain prompts. In some instances, ANI says, ChatGPT has falsely attributed statements to the news agency, hampering its credibility and misleading the public. Apart from seeking compensation for damages, ANI has asked the court to direct OpenAI to stop storing and using its work. In its response, OpenAI says it opposes the case being filed in India since the company and its servers are not located in the country and the chatbot has also not been trained there. In December, the Federation of Indian Publishers, which claims to represent 80% of Indian publishers including the Indian offices of Penguin Random House and Oxford University Press, filed an application in court saying that they were "directly affected" by this case and should be allowed to present their arguments as well. A month later, Digital News Publishers Association (DNPA), which represents leading Indian news outlets, and three other media outlets filed a similar application. They argued that while OpenAI had entered into licensing agreements with international news publishers such as the Associated Press and Financial Times, a similar model had not been followed in India. DNPA told the court the case would affect the livelihood of journalists and the country's entire news industry. OpenAI has, however, argued that chatbots are not a "substitute" for news subscriptions and are not used for such purposes. The court has not admitted these applications by the publishers yet and OpenAI has argued that the court should not hear them. But the judge clarified that even if these associations are allowed to argue, the court will restrict itself to ANI's claims since the other parties had not filed their own lawsuits. Meanwhile, OpenAI told the BBC it is engaging in "constructive partnerships and conversations" with news organizations around the world, including India, to "work collaboratively". Analysts say the lawsuits filed against ChatGPT across the world could bring into focus aspects of chatbots that have escaped scrutiny so far. Dr Sivaramakrishnan R Guruvayur, whose research focuses on responsible use of artificial intelligence, says that the data used to train chatbots is one such aspect. The ANI-OpenAI case will lead the court "to evaluate the data sources" of chatbots, he said. Governments across the world have been grappling with how to regulate AI. In 2023, Italy blocked ChatGPT saying that the chatbot's mass collection and storage of personal data raised privacy concerns. The European Union approved a law to regulate AI last year. The Indian government too has indicated plans to regulate AI. Before the 2024 election, the government issued an advisory that AI tools that were "under-testing" or "unreliable" should get government permission before launching. It also asked AI tools to not generate responses that are illegal in India or "threaten the integrity of the electoral process". — BBC
