A DPIIT-led committee has recommended that AI developers should get automatic access to copyrighted content, but only after paying mandatory royalties. The panel rejected a blanket TDM exception, insisting creators must be fairly compensated.
A committee set up by the Department for Promotion of Industry and Internal Trade (DPIIT) has released a working paper suggesting that companies developing Artificial Intelligence models must pay fair compensation to copyright holders for using their content to train Large Language Models (LLMs).
DPIIT, which functions under the Ministry of Commerce and Industry, had created this expert panel to study how copyright law should deal with the rapid growth of AI technology and whether the current legal system needs changes.
The committee, headed by DPIIT Additional Secretary Himani Pande and including several legal and technology experts, examined whether the Copyright Act, 1957 can handle the challenges brought by AI or if amendments are required.
A detailed working paper containing their findings and recommendations has been uploaded on the DPIIT website for public consultation.
Anyone wishing to submit suggestions or objections can email their comments to “ipr7-dipp@gov.in” within 30 days from December 8.
During the consultation phase, most stakeholders from the AI industry demanded a complete Text and Data Mining (TDM) exception so that AI systems could freely use copyrighted material for training.
On the other hand, representatives of the content industry were unanimous that only a voluntary licensing model would protect creators and ensure fair payment.
In its paper released on December 8, the panel rejected the demand for a blanket TDM exception. It stated that such a model would weaken copyright protection and leave creators without any control or compensation for the use of their work.
The paper clearly notes,
“It was not found to be a wise policy choice, especially for a country like India which has a rich cultural heritage and a growing content industry with immense potential.”
The committee also examined whether an opt-out mechanism for creators would work.
However, it found that this would still leave many smaller creators unprotected because they may not be aware of such options, may not have bargaining power, and may not be able to monitor whether their work is being used even after opting out.
The panel also warned that large numbers of opt-outs could reduce the availability of broad and diverse datasets needed for high-quality AI training.
Therefore, the committee concluded that creators should not be allowed to withhold their works from being used to train AI systems. It acknowledged that AI development requires access to large and high-quality datasets.
ALSO READ: Administrative Tribunals Occupy a Unique Space Between Executive and Judiciary: CJI Gavai
The report states that,
“With a majority view, the Committee decided to recommend a mandatory blanket license in favour of AI Developers for the use of all lawfully accessed copyright-protected works in the training of AI Systems, accompanied by a statutory remuneration right for the copyright holders.”
To ensure both easy access for AI developers and fair compensation for creators, the panel has proposed a hybrid model. Under this system, AI developers would get automatic rights to use any lawfully accessed copyrighted material, without negotiating individual licenses.
At the same time, creators would receive fair statutory royalties through a structured and simplified process.
The model aims to reduce compliance and transaction costs for AI developers, create a level playing field for all companies including startups, ensure easy royalty distribution, and reduce risks of AI bias or hallucination due to limited datasets.
The committee recommended creating a single central non-profit body, made up of associations of rights holders, which would be designated by the Central government under the Copyright Act.
This organization would collect payments from AI developers and distribute royalties to copyright owners. Only one representative organization per category of works would be allowed — either a registered copyright society under Section 33 of the Copyright Act, 1957, or a non-profit collective management organisation representing that class of works.
The panel’s report explains that,
“Certain percentage of the revenue generated from AI Systems trained on copyrighted content would be payable as royalties. The royalty rates would be fixed by a committee appointed by the government. By preserving the right of the copyright owners to receive royalties and administering it through a single umbrella organization made by the rights holders and designated by the movement, the model aims to provide an easy access to content for AI Developers for AI Training, simplify licensing procedures, reduce transaction costs, ensure fair compensation for rightsholders.”
The committee has suggested naming this body the ‘Copyright Royalties Collective for AI Training (CRCAT)’. Royalty distribution would be based on a Works Database, where creators would register their work.
AI developers would also need to provide a “sufficiently detailed” disclosure of their training datasets. They would have to share broad categories, nature and general sources of data used, but would not be required to reveal technical or confidential information.
The report mentions that developers must follow a basic rule of disclosure, where they provide a summary of the types of content used in training.
However, the National Association of Software and Service Companies (Nasscom) has strongly disagreed with the hybrid model.
It argued that rights holders should be given clear statutory protection against unrestricted Text and Data Mining (TDM) instead of imposing a mandatory blanket licensing model on the AI industry.
ALSO READ: ANI Before Delhi HC: “OpenAI Dilutes Market For Our Content, Causes Unfair Competition”
The committee included IAS officer Himani Pande, Simrat Kaur, Anurag Kumar, Advocates Ameet Datta and Adarsh Ramanujan, along with Raman Mittal, Chockalingam M, and Sudipto Banerjee.
The issue is also before the judiciary. The Delhi High Court is currently hearing a major lawsuit titled ANI Media Pvt. Ltd. v. OpenAI Inc., where ANI has accused OpenAI of copyright infringement.
ANI alleges that ChatGPT was trained on its copyrighted content without permission, making it the first such case in India. The court’s decision in this matter may influence how India approaches AI training and copyright enforcement in the future.
Read More Reports On AI Training
