“What Protects Our Music?”: How AI Exposes Copyright Gaps In India’s Folk Traditions

Over the past three years, IPRS has conducted copyright awareness camps across India, including in tribal and rural communities in Himachal, Rajasthan, Ladakh, Odisha, Jharkhand, Tamil Nadu, Bengal, the Northeast, and central India. The purpose was straightforward: to explain to musicians and performers how copyright works, what it protects, and how they can benefit from it.

What we did not anticipate was the question that came back to us again and again from folk and tribal musicians: What protects us?

When we explained that copyright vests in individual authors who create original works fixed in a tangible medium, the response was immediate and unsettling. Their music has no individual author. It has never been written down. It has been sung, played, and passed across generations within their communities for centuries. And under the Indian Copyright Act, 1957, none of that qualifies for protection.

This article is a response to what we heard in the field. It is not a theoretical exercise. It is an account of a structural problem that IPRS encountered through direct engagement with the communities whose music forms the bedrock of India’s cultural heritage, and an honest assessment of what can and cannot be done about it.

What Copyright Protects and What It Cannot

The Indian Copyright Act was designed for individual creators producing original works that are fixed in a tangible form. It provides a statutory monopoly for the author’s lifetime plus sixty years.

This framework has served composers, lyricists, and publishers well, and IPRS exists to administer these rights, ensuring that when music is used, its creators are compensated.

But traditional folk and tribal music is collectively authored, orally transmitted, and continuously evolved over centuries. It has no identifiable individual author, no date of creation, and no fixation in the sense the Act requires.

This is not a failure of implementation; it is a structural limitation of copyright as a legal framework. Folk and tribal music, by its very nature, falls outside the architecture of the Copyright Act.

The conundrum is stark: folk music deserves protection, but the legal tools designed for intellectual property are structurally incapable of providing it and the alternatives do not yet exist.

There is a further difficulty that must be stated honestly. When a contemporary composer takes a tribal melody and creates an arrangement, the copyright in that arrangement vests in the arranger, not in the community that originated the melody. The performer’s rights in a recording belong to the performer.

The producer’s rights belong to the producer. At no point does any right flow back to the source community. The existing copyright framework does not merely fail to protect traditional musicians; it can become the very mechanism through which their material is appropriated with legal cover.

Why AI Changes the Scale of the Problem

Artificial intelligence is now being trained on vast datasets of existing music, including traditional and folk music that has been digitised and made publicly accessible. AI systems can learn from a Baul song, a Rajasthani folk melody, or a tribal rhythm from the Northeast and generate new compositions derived from these sources at an industrial scale.

This is not a qualitatively new problem. Misappropriation of folk music is as old as the recording industry. Bollywood has borrowed from folk traditions for decades. But it is a quantitatively new problem, and the difference matters.

Earlier misappropriation required at least some human creative labour: an arrangement, a recording, a performance. AI can process and recombine folk material without human intermediary, at a speed and volume no individual composer could match.

The input is invisible, the process is opaque, and the output is monetisable. The absence of legal protection for the source material becomes far more consequential when extraction is automated.

For the communities we met in our camps, this is not an abstract concern. It is the prospect of their cultural heritage being absorbed into commercial systems from which they will never see a rupee.

Twenty-Five Years of International Negotiation and Still No Treaty

It would be misleading to present this problem as though it were newly discovered. The World Intellectual Property Organization (WIPO) established its Intergovernmental Committee on Intellectual Property and Genetic Resources, Traditional Knowledge and Folklore (IGC) in 2000.

For twenty-five years, the IGC has conducted text-based negotiations toward international legal instruments for the protection of traditional knowledge (TK) and traditional cultural expressions (TCEs), the formal term for what we commonly call folklore.

In May 2024, WIPO member states adopted the GRATK Treaty, the first WIPO treaty with provisions specifically for Indigenous Peoples and local communities. However, this treaty deals narrowly with patent disclosure requirements for genetic resources. It does not address traditional cultural expressions. No binding international instrument for the protection of TCEs currently exists.

India’s tribal musicians are living the consequences of this unfinished work. The international community recognises the problem, but it has not solved it.

What Is Actually Needed

Vague calls for “future-ready IP frameworks” will not serve the communities we heard from. What is needed are specific, actionable measures:

A sui generis legislative framework for traditional cultural expressions, separate from copyright. Copyright was not designed for collectively authored, orally transmitted cultural heritage, and it cannot be retrofitted to cover it. India should develop dedicated legislation drawing on the WIPO IGC draft articles and on models from countries that have attempted this, including Kenya, Panama, and the Philippines. This framework must recognise community custodianship as a distinct category of rights holder.

A national TCE registry linked to community custodians. India already operates the Traditional Knowledge Digital Library (TKDL), built primarily for patent-defensive purposes in the domain of traditional medicine. A parallel initiative for musical and performative TCEs would create a documented record linking traditional expressions to their originating communities, a precondition for any attribution or benefit-sharing system.

Mandatory disclosure requirements for AI systems trained on Indian cultural material. The EU AI Act already includes disclosure requirements for training data. India’s evolving Digital India framework should incorporate provisions requiring AI developers to disclose the use of traditional cultural material in training datasets. This is the minimum condition for transparency.

A benefit-sharing mechanism adapted for intangible cultural heritage. The Nagoya Protocol provides a model for benefit-sharing in the context of genetic resources. A parallel mechanism for TCEs would ensure that when commercial value is derived from traditional cultural expressions, a portion flows back to custodian communities. The operational questions, attributed to whom? distributed how?,are genuinely difficult, but they must be confronted rather than deferred.

Clarity on institutional responsibility. The legislative initiative must come from the Ministry of Culture and the Ministry of Commerce, which handles India’s engagement with WIPO,not from a collecting society. IPRS, with its existing infrastructure for rights registration, documentation, and royalty distribution, can serve as a technical partner in designing and administering such a system. But it cannot substitute for the sovereign authority required to create new categories of rights.

What We Owe the Field

IPRS went to the hinterlands and the remotest areas to teach copyright. What we learned instead was the boundary of our own mandate. We met musicians whose art predates our legal system by centuries and whose first question, What protects our music?, we could not answer.

We owe them more than an acknowledgement of the problem. We owe them a serious effort toward a solution, one grounded in legislation, informed by twenty-five years of international negotiation, honest about what copyright can and cannot do, and designed with the communities themselves at the table.

AI will continue to evolve. Its capacity to process and recombine cultural material will only grow. The question is not whether machines will learn from India’s folk traditions, they already are.

The question is whether we will build the legal and institutional architecture to ensure that the communities who created this music are visible, acknowledged, and compensated, or whether we will allow their heritage to be absorbed into commercial systems that return nothing to its sources.

The decisions we make now will determine whether that future is inclusive or extractive. The musicians we met in the field are waiting for an answer.

It is time we gave them one.