<\/p>\n
India’s proposed legislation requiring AI firms to pay royalties on copyrighted Indian content for training models marks a significant intervention in the global AI economy, raising questions about innovation, fair use, and international market dynamics.<\/p>\n<\/div>\n
India\u2019s proposal to make artificial-intelligence firms pay royalties for using copyrighted Indian content to train models marks a striking assertion of regulatory power by a country that has quickly become central to the global AI economy. According to the report by Rest of World, the draft framework would require tech giants to pay a mandatory blanket licence , a percentage of global revenue , and create a government\u2011designated collecting body to distribute payments to creators. [1]<\/a><\/sup><\/p>\n The Department for Promotion of Industry and Internal Trade\u2019s working paper lays out a so\u2011called \u201cOne Nation, One Licence, One Payment\u201d hybrid model that would give AI developers automatic access to all \u201clawfully accessed\u201d copyrighted works for training in return for statutory remuneration, while prohibiting creators from opting out. The proposal argues this single\u2011window approach will reduce transaction costs and litigation risk, and produce a steady revenue stream for India\u2019s vast and often informal creative sector. [2]<\/a><\/sup>[6]<\/a><\/sup><\/p>\n Officials behind the plan have made clear that royalties would be linked to firms\u2019 global AI revenues rather than what they earn in India, a choice that amplifies the financial stakes for multinational developers and underlines New Delhi\u2019s leverage as one of the world\u2019s largest markets for generative AI services. Proponents say the measure would bring legal certainty where courts and regulators have yet to define whether large\u2011scale training constitutes fair use. [4]<\/a><\/sup>[5]<\/a><\/sup><\/p>\n Industry bodies and some legal experts have pushed back. Nasscom has formally dissented, warning mandatory licensing could slow innovation, and private lawyers such as Rahul Matthan have described the payment model as \u201cdeeply flawed,\u201d arguing it risks concentrating proceeds with established rights\u2011holders while leaving small creators with little control or compensation. Matthan has urged a focus on downstream outputs , pursuing remedies when models reproduce substantial copyrighted passages , rather than blanket upstream licenses. [1]<\/a><\/sup><\/p>\n The committee behind the proposal reasons that litigation is expensive, slow and uncertain, pointing to high\u2011profile, drawn\u2011out disputes elsewhere as evidence that preemptive licensing would spare creators and companies protracted court battles. The paper recommends a centralised non\u2011profit collecting entity, membership for collective management organisations, and a government committee to fix royalty rates and oversee distribution to both registered and unregistered creators. [3]<\/a><\/sup>[6]<\/a><\/sup><\/p>\n Practical challenges remain significant. Critics note the difficulty of measuring how much any single work contributes to a model\u2019s output, the administrative capacity required to collect and equitably distribute royalties, and the risk that retroactive charges could hit nascent startups and smaller players disproportionately. The government counters that a mandatory blanket licence would lower compliance costs for startups by replacing countless individual negotiations with a single legal regime. [2]<\/a><\/sup>[6]<\/a><\/sup><\/p>\n The policy comes amid mounting global debate: Brazil\u2019s draft AI bill includes compensation for rights\u2011holders, Europe and the United States are grappling with transparency and fair\u2011use questions, and private settlements , such as a major authors\u2019 settlement reached in the U.S. , have already signalled that market pressures are shifting. India\u2019s proposal therefore represents one of the most interventionist regulatory experiments to date and could become a template for other countries seeking to monetise locally produced training data. [1]<\/a><\/sup>[5]<\/a><\/sup><\/p>\n If New Delhi follows through, tech companies that have already committed billions to India will face a clear choice: absorb licence costs into their deployment and product models or forgo one of their fastest\u2011growing markets. The government\u2019s approach underscores a broader tension between preserving incentives for creators and sustaining an open environment for AI innovation; how that balance is struck in India will be watched closely across capitals and courtrooms worldwide. [1]<\/a><\/sup>[3]<\/a><\/sup>[5]<\/a><\/sup><\/p>\nSource Reference Map<\/h3>\n