A committee under Department for Promotion of Industry and Internal Trade (DPIIT) has proposed a new regulatory framework titled “One Nation, One Licence, One Payment.” Under this model, AI firms — including global players — would require a mandatory blanket licence to use any lawfully-accessed copyrighted content for training their models.
Rather than negotiating with each rights holder individually, AI developers would pay royalties into a centralised fund or collective entity (tentatively called Copyright Royalties Collective for AI Training — CRCAT) which would then distribute payments to creators whose works are used.
Importantly, the payment obligation would apply even for works already used — the proposal includes retroactive royalties for past model-training on copyrighted Indian content.
📄 Why this matters — What pushed the government to act
As generative AI models like ChatGPT, Gemini and others have scaled globally, concerns have mounted over widespread use of copyrighted text, images, music, and video without compensation to creators. India’s creative industry and news organisations have been vocal about the lack of remuneration or consent.
The DPIIT committee argued that unrestricted access — under “text and data mining” exceptions or fair-use claims — can erode the incentives for human creators and harm the creative ecosystem in the long run.
By offering a blanket-licence model instead of opt-out or case-by-case licensing, the government aims to simplify compliance for AI firms while ensuring fair compensation for all rights holders, even independent or unregistered creators.
🔧 How the framework would work — Key features
- Blanket licence: Once obtained, AI firms can use “lawfully accessed” copyrighted works across categories (text, music, video) for training, without needing individual permissions.
- Central royalty fund (CRCAT): Payments go to a central body that pools royalties and distributes them among creators based on usage, rather than requiring direct deals.
- Statutory payment — not opt-out: Creators have no option to withhold their works; they must accept that if their content was lawfully accessible, it may be used, but they must be compensated.
- Retroactive coverage: AI models trained on Indian copyrighted content in the past — even before the law — could be liable for royalty payments.
- Transparency obligations: AI firms may have to disclose data used for training, enabling tracking of which works contributed to models — to determine royalty distribution.
✅ Potential impact — Who stands to benefit, who might be hit
For creators (authors, journalists, artists, musicians, filmmakers)
- Fair compensation for use of their work in AI training, even if content is publicly available.
- A simpler, centralised way to claim royalties without chasing multiple AI firms.
- Long-term income stream from AI-driven use of their creative output.
For AI companies (global and domestic)
- Legal clarity and a well-defined access route to Indian content corpus — reduces compliance risk and litigation.
- But potentially high costs: mandatory royalties (plus retroactive payments) may raise expenses significantly, especially for large models using massive datasets.
For the AI ecosystem & Indian digital content industry
- Could foster a more balanced ecosystem — encouraging responsible data use, and respecting creators’ rights.
- But might raise barriers for smaller AI startups or research labs due to increased costs, possibly tilting the balance toward large, well-funded firms.
⚠️ Concerns & Criticisms — What opponents say
Some industry groups — including NASSCOM — have criticised the proposal, calling it a “tax or levy on innovation.” They argue a blanket licence + mandatory payment regime may stifle innovation, raise costs for AI firms, discourage research, or push model-training outside India.
Critics also warn that retroactive payments could create huge liability for existing AI models — possibly leading to legal battles or outright withdrawal from the Indian market.
Smaller startups and research-driven AI labs may struggle under the financial burden, which could hamper open-source innovation and competition.
🌍 What this means globally — Precedent-setting law
If implemented, India’s “One Nation, One Licence, One Payment” model could become a global benchmark — offering a third way between the U.S. “fair use / text-and-data-mining exception” and the EU’s “opt-out from training” regimes.
Large global AI firms may need to adjust how they access data — licensing, royalty payments, more transparent sourcing. Other countries, especially in the Global South, could follow India’s lead to assert rights over their cultural and creative content.
It reflects a broader push to balance AI innovation with cultural and creative rights — ensuring that content-creators — not just big tech — benefit from AI-driven value.
🕒 What to watch — What happens next
- The proposal is currently in public consultation — AI firms, publishers, creators and civil-society groups have until end of the comment period (30 days) to respond.
- Based on responses, the government may finalise the legislation — defining royalty rates, structure of the central collective (CRCAT), and mechanisms for tracking usage and distributing payments.
- Legal challenges are likely — global firms may push back, citing global precedent, technicality of data ingestion pipelines, and possible impacts on model-training operations.
- Industry reaction: how many firms continue investing in India vs rerouting training infrastructure to other geographies to avoid royalties.
💡 Final thought
With the “One Nation, One Licence, One Payment” proposal, India is taking a bold — perhaps historic — step to reshape the balance between AI innovation and copyright protection. If implemented, it could establish a fairer, legally structured system where creators are compensated, and AI firms continue building models — but under responsibility and transparency.
Whether this becomes a global model or a cautionary tale will depend on how stakeholders respond, how costs and benefits balance out, and how the legislation is implemented.
