India has moved to the center of the global debate over AI training data with a sweeping proposal that would force companies such as OpenAI and Google to pay royalties for using copyrighted material — a move that could reshape how the world’s largest AI firms operate inside one of their most important growth markets.
On Tuesday, the country’s Department for Promotion of Industry and Internal Trade released a proposed framework that would grant AI developers automatic access to all copyrighted works for training in exchange for paying royalties into a new collecting body composed of rights-holding organizations. According to Tech Crunch, the money would then be distributed to creators.
India’s proposal argues that such a “mandatory blanket license” would lower compliance costs for AI firms while ensuring that writers, musicians, artists, and other creators are compensated when their work is scraped into datasets that power commercial models.
The push lands at a moment when concerns over unlicensed training have erupted across global markets. Authors, newsrooms, artists, and photographers in the U.S. and Europe are taking AI developers to court, adding pressure on regulators to determine whether training qualifies as fair use. The legal uncertainty has not stopped AI companies from expanding, but it has left them exposed to costly litigation. India, by contrast, is offering one of the most interventionist solutions seen anywhere, essentially creating a statutory license that both guarantees compensation and standardizes access.
At the heart of the proposal is an eight-member government committee formed in April. In a 125-page submission, the group argues that a blanket license provides “easy access to content for AI developers,” reduces transaction costs, and ensures that rightsholders are paid. The committee describes the new collecting body as a “single window,” eliminating the need for individual negotiations and enabling royalties to flow to both registered and unregistered creators.
India’s pitch reflects the country’s surging importance in the global AI economy. OpenAI CEO Sam Altman has said India is the company’s second-largest market after the U.S. and may ultimately become its largest. The committee notes that AI companies generate significant revenue from Indian users while also drawing heavily on creative works from the Indian ecosystem, a dynamic it says justifies a “balanced framework” that returns some of that value to creators from the outset.
India is now positioning itself as the first major jurisdiction to impose a structured royalty system on the entire AI training pipeline. Its model goes far beyond transparency requirements debated in the U.S. and the European Union, where policymakers are still wrestling with disclosure obligations, audit mechanisms, and whether training can be considered transformative.
The solution, for India, is to dispense with those ambiguities and create a mandatory mechanism that pays creators while giving AI developers immediate legal certainty.
But the reaction from the technology industry has been anything but unified. Nasscom, which represents major technology companies including Google and Microsoft, filed a formal dissent urging India to adopt a broad text-and-data-mining exception instead. It argued that AI developers should be free to train on copyrighted content so long as they access the material lawfully. Nasscom warned that mandatory licensing could slow innovation and said rights holders who object should be allowed to opt out, rather than forcing companies to pay for all training data.
The Business Software Alliance, whose members include Adobe, Amazon Web Services, and Microsoft, also urged the government to avoid relying solely on licensing. In its submissions, BSA pushed for an explicit text-and-data-mining exception, arguing that a mandatory licensing regime might be impractical and could limit training datasets so severely that model quality suffers. It warned that restricting models to licensed or public-domain material could “increase the risk that outputs simply reflect the trends and biases” of limited datasets.
The committee, however, rejected the idea of combining a broad exception with an opt-out, arguing that such systems either weaken copyright protections or create enforcement challenges so enormous that rights holders cannot realistically protect their work. Instead, the proposed hybrid model would give AI developers immediate, automatic access to all copyrighted works — but require them to pay royalties into the central collecting body.
India is the world’s largest market by population, one of the fastest-growing digital economies, and an increasingly critical testing ground for enterprise AI adoption. If India adopts the system as proposed, every AI developer operating in the country would face predictable royalty obligations but also gain unrestricted access to rich, culturally diverse datasets. Companies may ultimately accept that trade-off if it provides legal certainty missing in Western markets. But royalty payments could reshape cost structures, especially for large-scale model training that depends on vast, continuously updated datasets.
For creators, the proposal marks a rare moment when regulatory momentum sits squarely on their side. Indian courts are already reviewing cases that test whether training violates copyright law. A Delhi High Court case brought by ANI against OpenAI has pushed judges to interrogate whether training itself counts as reproduction. Globally, similar lawsuits are moving through U.S. federal courts and European jurisdictions. India’s framework would bypass years of legal uncertainty by requiring compensation up front.
The political momentum behind the plan reflects India’s ambitions to regulate AI on its own terms rather than follow frameworks built in Washington or Brussels. The country sees AI as a strategic economic pillar and appears intent on designing a copyright system that reflects the size of its creative industries and the scale of its digital markets.
The government has opened the proposal for public consultation, giving companies and stakeholders 30 days to submit comments. After reviewing submissions, the committee will finalize its recommendations before the framework advances through the government.
India’s move sets up a dramatic global test case. If implemented, its system could become a template for other emerging markets with large creator communities but limited bargaining power against Big Tech. It could pressure Western regulators to accelerate their own debates. It could force companies like OpenAI, Google, and Microsoft to operate with new royalty liabilities in one of the world’s most consequential markets.
Or it could spark the industry to mount an aggressive pushback phase, warning that a licensing-heavy regime would slow innovation and limit India’s competitiveness.
Either way, India has just injected one of the boldest, most far-reaching proposals yet into the debate over who gets paid in the age of AI.
