India has moved to the forefront of global AI governance by proposing a groundbreaking "One Nation One License" copyright framework for artificial intelligence training data. This proposal, reported in December 2025, would establish a mandatory blanket licensing system under which AI companies must obtain a single national licence to train their models on copyrighted content, with statutory royalties paid to rights holders through a new body: the Copyright Royalty Collection and Administration Tribunal (CRCAT).

The proposal directly addresses one of the most contentious global debates in AI governance: whether large language models and other generative AI systems that train on internet-scraped copyrighted content — books, articles, music, artworks — are engaging in infringement. Major cases are pending globally, including lawsuits against OpenAI and Google by publishers and the New York Times.

India's model takes a middle path: it neither bans AI training on copyrighted content nor allows it for free. Instead, it creates a statutory licensing framework similar to compulsory licensing in the pharmaceutical sector, where use is permitted but rights holders must be compensated through a fixed royalty rate determined by the CRCAT.

If enacted, this would make India the first country to implement formal legislation specifically governing generative AI copyright — ahead of the European Union's AI Act (which addresses risk categories but not training data royalties comprehensively) and proposed US legislation.

For UPSC/RAS aspirants: this connects to intellectual property rights (GS Paper II/III), digital economy governance, India's AI strategy (IndiaAI Mission), and the creative industries. It also raises questions about Article 19(1)(a) (freedom of expression), the Copyright Act 1957, and India's obligations under TRIPS and the Berne Convention.