In one its first major regulatory move towards AI, India has come up with a new government proposal that would require companies like OpenAI, Google, Anthropic and others to pay royalties for using Indian copyrighted content. The draft, prepared by the Department for Promotion of Industry and Internal Trade, outlines a statutory licence that would unify permissions for all legally accessed copyrighted works and attach compulsory royalties, including for past training cycles.
This, if enacted, would be one of the first major AI regulations globally by a major tech economy, that focuses on use of copyrighted material by AI companies.
Along with this, India’s draft framework says it would require large model developers to disclose what data they used, signaling a major shift toward transparency and monetization in one of the world’s largest content markets.
The proposal forms part of a wider push by India to assert control over its creative economy at a time when AI models are consuming unprecedented volumes of text, audio, images, and video. Under the envisioned system, companies training AI in the country would no longer negotiate private deals with individual creators and publishers. Instead, they would operate under a single national licence that gives them access to copyrighted content – but makes royalty payments mandatory. This is clearly different from the ‘opt-out’ approach common in the US and parts of Europe, where creators must actively request that their work be excluded from AI training datasets rather than receiving any guaranteed compensation.
The government proposes creating a centralized authority to manage the system, tentatively described as the Copyright Royalties Collective for AI Training (CRCAT). This body would be responsible for collecting payments from AI companies, tracking usage, and distributing funds to creators across sectors – including authors, journalists, filmmakers, composers, digital artists and others.
The royalty rates themselves would be decided by a dedicated committee, with revisions expected periodically to reflect market dynamics. According to the draft, rates may be linked to global revenues earned from AI products that rely on Indian content during training.
One of the proposal’s most consequential elements is its retroactive scope, meaning AI models that are already trained and commercially deployed could still face royalty liabilities if they used Indian copyrighted material. The draft also introduces strict transparency rules requiring companies to disclose what data they trained on, where it came from, and whether it was paywalled or publicly accessible. The government has opened the proposal for public feedback for 30 days, after which policymakers will consider revisions before deciding whether to turn the draft into a binding law.
However, the proposal has already sparked concerns that mandatory royalties tied to global revenues could discourage companies from training models in India and introducing their most advanced systems. There is also worry that smaller startups may struggle with the added costs, leaving the field dominated by larger firms, along with a strong possibility for fresh controversies around whether licensing terms should be set by the market or the government.
The Tech Portal is published by Blue Box Media Private Limited. Our investors have no influence over our reporting. Read our full Ownership and Funding Disclosure →
