The Indian government is preparing to change copyright laws to meet the challenges of artificial intelligence within the next three years, a senior official said on December 11, 2025. The Department for Promotion of Industry and Internal Trade (DPIIT) recently released a working paper on AI and copyright, suggesting a "blanket licensing" system. Under this, websites whose data is used by large language models (LLMs) like ChatGPT would get royalty payments through a copyright society. This society would then share money with content publishers such as book and news organisations. This plan aims to settle conflicts between online content creators and AI firms that train their models using large amounts of internet-scraped text. Publishers want compensation for their role in developing these AI tools. The DPIIT proposal allows AI firms to scrape content legally but requires payments after their models go commercial. In about two months, DPIIT will release another paper exploring if AI-generated content can be copyrighted and how authorship is assigned, said DPIIT Additional Secretary Himani Pande. After this, the government will likely propose amendments to the Copyright Act, 1957, in Parliament. Ms. Pande explained that the Copyright Royalties Collective for AI Training (CRCAT) would only demand payments from AI firms once they start selling their models, not during data collection for training. This issue has caused global legal battles. News agencies like ANI and The New York Times have sued OpenAI, creator of ChatGPT, for using their content without permission. OpenAI denies these claims. However, the tech industry body Nasscom, representing firms like Google and Meta, disagrees with DPIIT's approach. Nasscom wants publishers to have the choice to opt out of data being used for AI training. They warn that forced blanket licensing could lead to more disputes. A major tech firm noted concerns that DPIIT's plan shifts the legal burden unfairly onto AI developers, who would need to prove they did not use copyrighted content. This is difficult because AI tools work with probabilities, not exact copying. Ms. Pande said the government will consider industry feedback as consultations continue.