The Indian government is working on a new framework to decide how AI companies can use copyrighted content for training. The government has released the first part of a working paper that looks at how copyright law should deal with generative AI. This is an important step because AI models are trained on huge datasets that include books, news articles, music, artwork, and more. Until now, there has been no clear rule on whether creators should be paid for this.
The Department for Promotion of Industry and Internal Trade formed an eight-member committee in April 2025 to study the issue. They proposed a new hybrid licensing system that could directly affect companies like OpenAI and Google.
The proposal gives AI developers automatic access to all lawfully available copyrighted works for training. They do not need to negotiate with individual creators. This is meant to keep AI research smooth and avoid delays. But once an AI tool becomes a commercial product, the company will have to pay royalties. These payments will be collected by a central body and distributed to creators. The royalty rates will be set by a government-appointed committee and can be challenged in court.
The committee rejected the idea of free access without payment. It said that would reduce incentives for human creativity and lead to fewer original works in the long run.
This framework matters because India is a big market for AI. This is the reason all major AI companies are offering free access to their premium models for a limited time. OpenAI CEO Sam Altman has already confirmed that India is already their second-largest market and is growing fast to become the biggest market.
The committee noted that AI companies are earning from the content created for Indian publishers and creators. So, they should also return some value to the publishers and creators who own the content.
India’s proposal came at a time when AI companies are facing legal challenges around the world for using copyrighted data without permission. Even some companies in India are also fighting legal cases against AI companies for using their data to train AI models. Since there is no clear legal answer to how this matter should be tackled, India wants to create its own system to create a clear guideline.
This approach feels practical. AI companies are not only using other people’s data to train their models, but they are also affecting the revenue of creators and publishers. For example, publishers like us are seeing less traffic because users now get information directly from AI overviews. But the information shown in those AI answers still comes from websites like ours. We have already seen several big publications announce layoffs or shut down because of this shift. A central licensing system can help both sides, but only if it is implemented properly.
