“One Nation, One Licence, One Payment”: What the DPIIT’s AI Copyright Proposal Means

Every generative AI model is trained on enormous amounts of text, images, music and video, much of it protected by copyright. Who should be paid for that, and how? Around the world, the answer is being fought out in courtrooms. India is trying something different: a statutory licence.

In December 2025, a committee set up by the Department for Promotion of Industry and Internal Trade (DPIIT) published a working paper with an ambitious title, “One Nation, One Licence, One Payment”. This guide explains what the DPIIT AI copyright working paper proposes, why it has drawn both support and fierce criticism, what it would mean for creators and AI startups, and what comes next.

Where the DPIIT AI copyright working paper came from

DPIIT formed an eight-member committee on 28 April 2025, headed by Additional Secretary Himani Pande, to identify the issues raised by AI systems, examine the existing legal framework, assess its adequacy and recommend changes. On 8 December 2025, the committee issued Part I of its working paper on generative AI and copyright.

Part I deals only with the input side: the use of copyright-protected works as training data. Part II, still to be published, is meant to cover the copyrightability and authorship of AI-generated outputs, including moral rights and liability for infringing outputs. Public comments on Part I were initially due by 7 January 2026, and the deadline was extended to 6 February 2026.

What the “hybrid model” actually proposes

At the heart of the DPIIT AI copyright working paper is a mandatory blanket licence. AI developers would be permitted to use all lawfully accessed copyrighted works for training, coupled with a statutory right to remuneration for copyright holders. Instead of AI companies negotiating millions of separate licences, the law would grant a single licence and channel payments to rights holders.

Royalties would be triggered only when the AI product earns revenue, so free research and pre-revenue development wouldn’t immediately cost money. Payments would flow through a central collecting body, described in the proposal as the Copyright Royalties Collective for AI Training (CRCAT). The paper also proposes a rate-setting committee, mandatory registration systems and grievance redressal mechanisms. The working paper doesn’t fix the royalty rate itself, which is one of the biggest open questions.

The proposal wasn’t unanimous: there was dissent inside the committee. It is a working paper for consultation, not a Bill, and nothing in it is yet law.

The criticism: an “AI tax” in the name of copyright?

The DPIIT AI copyright working paper has drawn strong criticism from several directions. One line of critique argues that the paper builds an elaborate regulatory apparatus without first establishing that AI training infringes copyright under the Copyright Act, 1957 at all. If training is not infringement, or is fair dealing, critics ask why AI developers should pay a statutory royalty for it. We examine that question in our guide to ANI v. OpenAI and fair dealing.

Others see the remuneration scheme as effectively a tax on AI development dressed up as copyright. Professor Arul George Scaria’s analysis, published in May 2026, frames exactly that question. On the other side, many creators and publishers welcome the idea that they’d be paid without having to sue each AI company individually, although some object that a mandatory licence takes away their right to say no.

There’s also an important difference from Europe. Unlike the EU regime, which gives creators an opt-out from AI training datasets, Indian copyright owners currently have no statutory right to prevent their works from being used in AI training. A mandatory licence would make that position permanent, in exchange for payment. For Indian companies serving European customers, see our guide to the EU AI Act.

What it means for creators and AI startups

Whether you are a creator or an AI developer, the DPIIT AI copyright working paper is worth reading in full, because the details of registration, rate-setting and distribution will decide who actually benefits. For creators, including writers, artists, musicians, photographers and publishers, the practical advice is to prepare for a world where registration and records determine who gets paid. Keep clear records of what you own, consider registering important works with the Copyright Office, and follow how any collecting body would identify and distribute royalties. Your publisher or label contracts may also need to address who receives AI training royalties. For ownership of AI outputs themselves, see our guide to AI-generated content and copyright.

For AI startups, the hybrid model would bring legal certainty at a price. Lawful access becomes the key condition, and the working paper proposes that the lawful-access requirement attach prospectively only. Startups should document where their training data comes from and how it was accessed, model the possible royalty cost against revenue, and avoid training on pirated or paywalled content obtained unlawfully, which would fall outside any statutory licence. Inside the business, set clear rules for what employees may feed into AI tools; see AI tools at work and confidentiality.

There’s also a timing point. Because the DPIIT working paper is only a consultation document, nobody has to pay a statutory AI training royalty today. But AI startups raising money in 2026 are already being asked by investors how a statutory licence would affect their margins, and creators’ organisations are lobbying for higher rates. Startups that have documented lawful access to their data and modelled a royalty scenario will answer those questions far more convincingly than those that haven’t. The proposal also interacts with litigation. In ANI v. OpenAI, the Delhi High Court is considering whether AI training is fair dealing under Section 52. If India enacts a hybrid licensing model first, the statute may overtake the litigation, and the outcome of that case may in turn influence how the government shapes the law.

What comes next, and quick answers

If you want to influence the outcome, stay engaged: DPIIT has consulted publicly on Part I, and further consultation on the DPIIT AI copyright working paper and any draft Bill is likely. The next milestones are Part II of the working paper, on the copyrightability, authorship and moral rights of AI-generated outputs, and any draft amendment to the Copyright Act. A DPIIT official said in December 2025 that Part II was expected in about two months, but as of the latest reports it had not yet appeared. Until legislation is passed, the existing Copyright Act applies, and creators retain their ordinary remedies for infringement: civil suits for injunctions and damages under Section 55, criminal complaints under Section 63, and takedowns of infringing outputs on platforms. Realistic synthetic output raises its own duties; see our guides on deepfakes and personality rights and labelling AI-generated content.

Quick answers

What is “One Nation, One Licence, One Payment”?

The title of DPIIT’s December 2025 working paper, which proposes a mandatory blanket licence for AI training on lawfully accessed copyrighted works, with statutory royalties paid to rights holders.

Is the DPIIT AI copyright proposal law?

No. It is a working paper for consultation. Any change would require an amendment to the Copyright Act.

Can Indian creators opt out of AI training?

Not under current law. India has no statutory opt-out, and the mandatory licence model would not add one.

Related reading: AI copyright in India vs the US; also AI training data licensing with Indian publishers; also copyright in legal headnotes and AI training and structuring AI training data licences after Ross; also Getty v. Stability AI at the Court of Appeal.

Final word

At Halverton & Co., we work with founders, startups and technology-driven businesses on AI and copyright, content licensing, AI startup compliance and infringement disputes. We practise in Jharkhand, Maharashtra and before the Supreme Court of India, and act as fractional legal counsel for companies that need senior legal support without a full in-house legal team. Halverton & Co. is built on a simple idea: Where tech needs law! If this issue affects your business, write to us at office@halvertonandco.com, or get in touch, and we’ll help you work out where you stand.

This article reflects the law and developments reported up to early October 2026. It is for general information only and is not legal advice. Please take advice on your specific facts before acting.