Summary
In December 2025, a DPIIT constituted committee released Part One of its Working Paper on Generative Artificial Intelligence and Copyright, proposing a hybrid licensing framework under the caption One Nation One License One Payment. The model seeks to balance AI innovation and copyright protection through a mandatory statutory licensing mechanism and centralised royalty collection and distribution. This article examines the assumptions underlying the proposed framework and evaluates whether they reflect the realities of authorship, copyright ownership, collective management, and AI development in India. It argues that several foundational assumptions require reconsideration and that the proposed model may not meaningfully serve the interests of authors and creators or support sustainable AI innovation. The article suggests that a clearly defined AI training exception, coupled with a statutory right vested in authors and creators, may offer a more workable and balanced approach for India.
In December 2025, the committee constituted by the Department for Promotion of Industry and Internal Trade DPIIT Government of India released Part One of its Working Paper on Generative Artificial Intelligence and Copyright. This part of the working paper focused on the use of copyrighted works for training and related purposes and proposed a hybrid licensing model under the caption One Nation One License One Payment as the most suitable option for India to balance AI innovation and copyright protection.
In simple terms, the proposed model envisages a mandatory statutory licensing scheme through the establishment of a Copyright Royalties Collective for AI Training CRCAT. The CRCAT is expected to collect royalties on behalf of copyright owners for the use of copyrighted works in AI training and related activities. Copyright owners will be granted a statutory right to royalty, and a Government committee will determine royalty rates based on the revenues of AI companies. The royalties collected by CRCAT are to be distributed to copyright societies or collective management organisations CMOs, which will then redistribute the amounts to copyright owners. Under the proposed model, copyright owners will not have the option to opt out of the blanket licence. Further, where authors of particular categories of works do not have a functioning CMO, they may not receive any royalty if the relevant CMO is not constituted.
Assumptions Underlying the Licensing model The committee has proposed this hybrid licensing model as a mechanism to balance the interests of human creativity and the progress of AI innovation in India. However, the model rests on several assumptions that require closer examination. On a careful reading of the working paper, these assumptions do not appear to be supported by evidence or by the realities of how copyright operates in practice.
Assumption One
“To protect and reward human creativity through the copyright system, copyright owners must be protected from the use of their works by AI businesses.”
The premise that copyright protection directly promotes human creativity remains unproven. Most authors create works independent of copyright incentives and protection possibilities. Creation precedes copyright, not the other way around. Even if copyright is assumed to play an incentive role, the economic benefits arising from copyright rarely accrue to authors and creators.
In sectors such as publishing, music, cinema, and the arts, copyrights are largely owned by producers, publishers, record labels, and other commercial entities. It is well recognised that the revenues generated from these copyrights do not sufficiently flow to authors and creators. This imbalance was acknowledged by the legislature in 2012, when statutory royalty rights were introduced for authors of literary and musical works incorporated in films and recorded music, which required producers to share royalties with authors.
If this fundamental assumption is not addressed, the proposed hybrid licensing model is likely to result in additional revenues for copyright owning entities, with little benefit reaching authors and creators. This outcome would undermine the stated objective of balancing human creativity and AI innovation. If new rights are to be introduced in the context of AI, they should vest in authors and creators rather than in copyright owners, possibly through the creation of a specific right relating to AI use under different categories of copyrighted works.
Assumption Two
“A text and data mining exception is the only possible copyright exception for AI training, and existing exceptions do not address AI training issues.”
In its working paper, the committee reviewed text and data mining exceptions in other jurisdictions and concluded that they are unsuitable because they do not cover certain types of AI training uses. However, the working paper does not explain why these uses cannot be addressed through a suitably drafted copyright exception under Indian law. A well defined AI training exception, with appropriate limitations, could be a viable alternative. Such an exception, when combined with a specific right for authors and creators, could support AI innovation while protecting creative interests, without the need for a complex and administratively burdensome licensing framework.
In the paper, the committee also came to the conclusion that the fair dealing provision under Indian copyright law does not cover AI training. However, the Copyright Act contains a wide range of exceptions, and different exceptions may apply depending on the nature and context of the use for AI. The assumption that fair dealing or fair use are the only relevant exceptions is therefore misplaced. Also, whether a particular use amounts to fair dealing cannot be predetermined in the abstract for all forms of AI training.
In the said context, this aspect may be reviewed before assuming that other viable exceptions do not exist. Also, setting aside the assumption, a copyright exception specifically for AI purposes may be carefully considered before implementing the proposed hybrid license model.
Assumption Three
“CMOs and copyright societies have the experience and capacity to manage royalty collection and redistribution efficiently.”
The history of CMOs and copyright societies in India has been marked by persistent concerns relating to governance, transparency, and efficiency. Allegations of misuse of funds, administrative inefficiencies, and limited accountability have been common. While some CMOs function reasonably well, the overall ecosystem does not inspire confidence among authors, creators, or even copyright owners.
Additionally, most CMOs continue to be dominated by copyright owning entities, and though authors in certain sectors may be members, the benefits flowing to them are often limited. In this context, the assumption that CMOs can effectively administer a large scale AI royalty regime rests on weak foundations. Alternatives that allow royalties to flow directly to authors and creators, similar to direct benefit transfer mechanisms, merit serious consideration.
Assumption Four
“Government control, regulation, and oversight are the most effective means of balancing AI innovation and copyright protection.”
The committee has proposed the establishment of CRCAT and a Government appointed royalty setting committee to determine royalty rates, oversee distribution mechanisms, and regulate the use of copyrighted works by AI companies. While regulatory oversight has a role to play, extensive Government control may introduce administrative delays and procedural complexities.
Experience with existing intellectual property institutions in India shows that administrative and quality problems are difficult to resolve. For example, despite multiple reform efforts at the Indian IP Office, challenges relating to processes and quality of decision making continue to subsist. In a rapidly evolving technological environment, excessive regulation may impede innovation rather than facilitate it. Therefore, a framework with limited Government intervention and clearly defined statutory principles may be better suited to balancing AI innovation with the interests of authors and creators.
Assumption Five
“The copyright driven entertainment industry requires protection from AI use of copyrighted works.”
The copyright owners identified by the committee, including producers and entertainment companies, represent only a small segment of India’s creative ecosystem. The committee itself recognises that creators and authors come from diverse backgrounds across the country, and the entertainment industry accounts for a relatively small proportion of the total volume of creative works produced. Therefore, giving special importance to the entertainment industry that holds only a small portion of creative works may not be justified.
Also, in today’s context, every individual using a social media platform is a creator. With the increasing use of AI tools, the volume of works created by individuals on a daily basis is growing rapidly, and these works also form part of the training pool for AI systems. If policy responses focus primarily on protecting established entertainment content owners, the proposed model risks reinforcing existing revenue structures rather than addressing the interests of the broader creator community.
Further, the number of works created using AI tools is increasing exponentially, and it is likely that AI generated works will soon constitute a substantial proportion of creative output. In such a context, the relative importance of traditional copyright owners may change, and existing copyright incentive structures may lose relevance. Any policy framework should take this evolving reality into account, and proposing a policy/model based only on the present state of affairs and one/two sectors may not yield the balance sought.
AI Training Exception and Rights of Authors and Creators
When viewed against these assumptions, a clearly defined statutory exception relating to AI training, subject to specific limitations, appears better suited to India’s needs. Unless AI companies are able to access copyrighted works without excessive administrative hurdles, AI innovation in India is unlikely to progress meaningfully.
At the same time, the interests of human creativity can be protected by granting authors and creators a specific statutory right relating to AI use of their works. Strengthening the position of authors and creators, rather than further entrenching the rights of copyright owners, is more likely to achieve the balance of AI innovation and creative endeavour the committee seeks.
As an author of several published books and several articles, and as an intellectual property practitioner, the revenue prospects offered by the proposed hybrid licensing model do not appear compelling to me. There is a real risk that the model will not materially benefit authors and creators, while substantially strengthening the position of copyright owners. Such an outcome would not be conducive to the long term development of copyright law or to the protection of human authorship and creativity in India.
“In a nation unified by diverse forms of expression and marked by immense AI potential, an open and liberal exception, supported by a well defined statutory right for authors and creators, offers a more credible path to balancing AI progress with creative endeavour.”