India has unveiled a sweeping proposal to regulate how generative artificial intelligence (GenAI) systems use copyrighted material, recommending a mandatory blanket licensing framework that would allow developers to train AI models on any lawfully accessed copyrighted content while ensuring compensation for creators. The proposal appears in a new 125-page working paper released in December 2025 by the Department for Promotion of Industry and Internal Trade (DPIIT).
The paper, Working Paper on Generative AI and Copyright – Part 1, marks the country’s most detailed attempt yet to balance AI innovation with the protection of creative industries. It follows months of consultations with technology companies, collective management organisations, media groups, and other rights holders.
DPIIT’s committee emphasized that GenAI’s development depends on access to vast quantities of high-quality data, including copyrighted works—but warned that unlicensed training threatens to undermine economic incentives for human creators. The report cites concerns that continued uncompensated scraping of creative works could “weaken the creative ecosystem,” particularly for small and underrepresented artists whose livelihoods rely on copyright protections.
After reviewing global approaches—including U.S. fair use interpretations, Japan’s broad text-and-data-mining exception, and Europe’s rights-holder opt-outs—the committee concluded that existing copyright law does not adequately address AI training. It rejected a broad text-and-data-mining exception for commercial AI systems, arguing that it would “leave human creators powerless” and disproportionately benefit large tech companies. Even opt-out mechanisms, the paper notes, are impractical for most creators and ineffective once data is scraped, stripped of metadata, and reused downstream.
Instead, the committee proposed a hybrid model built around a statutory blanket license. Under the framework, AI developers would automatically receive the right to train models on copyrighted materials, while rights holders would receive royalties collected and distributed by a new centralized, non-profit entity—the Copyright Royalties Collective for AI Training (CRCAT). The government would set royalty rates, subject to judicial review. Creators, including non-members of copyright societies, would register their works to receive payments.
The paper argues the model reduces transaction costs for startups, ensures predictable access to training data, and mitigates harmful model bias by enabling more diverse datasets. At the same time, it preserves what the report calls the “foundational pillar” of India’s creative economy by establishing a guaranteed compensation mechanism for artists, musicians, writers, and other rights holders.
Industry group Nasscom dissented, reiterating its support for a TDM exception with machine-readable opt-outs. However, the committee maintained that only a statutory licensing approach can achieve long-term balance and legal certainty across India’s expanding AI ecosystem.
The working paper is now open for stakeholder feedback and will inform India’s broader AI governance agenda, including forthcoming guidance on the copyright status of AI-generated outputs.
Need Help?
If you have questions or concerns about any global guidelines, regulations and laws, don’t hesitate to reach out to BABL AI. Their Audit Experts can offer valuable insight, and ensure you’re informed and compliant.
