India’s leading IT industry body, Nasscom, along with several major technology companies, has formally opposed a new proposal from the Department for Promotion of Industry and Internal Trade (DPIIT) that would mandate a blanket licence for artificial intelligence (AI) training on copyrighted content. The pushback comes just two days after the government released its working paper outlining a hybrid model for AI training that requires companies to pay creators while preventing rights holders from opting out of the training process.

The DPIIT committee, formed in April, had the support of most of its members for the proposed model. However, Nasscom, which has two representatives on the committee, has strongly dissented. The trade body argues that the proposal—as currently drafted—is technically unworkable, economically unsustainable and likely to be nearly impossible to enforce. Although DPIIT initially planned to include only a brief summary of Nasscom’s objections, the industry body insisted on the full 24-page dissent being incorporated into the final report, and it is now appended as an annexure.

Nasscom’s Core Arguments: Technical and Economic Impossibility

Nasscom’s primary recommendation is that India should allow Text and Data Mining (TDM) for AI training as long as the content is accessed legally. It stated that creators must be given the option to prevent their content from being used—either through a machine-readable opt-out mechanism or contractual clauses.

A person closely linked to Nasscom explained that general-purpose AI platforms—including large language models such as ChatGPT—operate on drastically reduced margins because token prices have collapsed by nearly 95% since 2023. The person added that none of the major GPT model developers are currently profitable. In such a landscape, a mandatory royalty regime imposed by the government would be extremely difficult for AI companies to sustain.

Under the proposed DPIIT model, royalty payments would be made to a central body called the Copyright Royalties Collective for AI Training (CRCAT). This organisation would be responsible for collecting fees—likely based on company revenue or scale—and distributing payments to creators, authors, publishers, artists and other rightsholders. AI companies would not negotiate directly with creators; rather, a single government-appointed committee would determine the applicable rates.

Why Big Tech Is Opposed

Technology firms argue that the proposal fails to reflect how AI models work. Large language models (LLMs), they note, transform text into numerical tokens and do not store copyrighted content in its original form. Their outputs are also non-deterministic, meaning the same prompt can produce different results each time. This, they argue, makes it nearly impossible to trace or attribute specific copyrighted material or justify royalty payments tied to specific content.

Moreover, experts warn that developers can further reduce traceability by adjusting model parameters—adding noise, increasing randomness, or other techniques. Some of these methods are already being used globally in response to copyright-related lawsuits. These technical realities, they say, raise serious doubts about whether a royalty-based licensing system could be implemented or monitored effectively.

Tech giants like Meta and Google are also critical of the proposal because it would increase their costs significantly and complicate AI training processes. One industry executive told HT that AI helps creators by lowering production costs and enabling more content creation. They argued that India’s rejection of TDM exceptions runs counter to practices in the EU, Japan and Singapore, all of which permit TDM with certain safeguards. The executive further warned that mandatory blanket licensing could slow AI innovation, hinder startups, and expose companies to lengthy legal challenges.

Impact on Startups: A Major Concern

Tech policy experts argue that the DPIIT proposal could be particularly damaging to early-stage companies. Jameela Sahiba, associate director at The Dialogue, said the move significantly increases compliance burdens and fixed costs for young firms that operate with limited resources and irregular revenues. Even minor royalty obligations, she noted, could overwhelm startups’ budgets, potentially shutting out new entrants from India’s AI ecosystem.

Questions Around Committee Composition and Process

Industry insiders also believe that the structure and composition of the DPIIT committee influenced the outcome. A big tech executive alleged that the recommendations were bound to favour publishers because two of the committee members—attorneys Adarsh Ramanujan and Ameet Datta—are involved in the ongoing OpenAI vs ANI case on behalf of news publishers.

HT previously reported that one committee member had requested removal, citing a lack of expertise in AI or copyright, but remained on the panel despite seeking withdrawal.

Conflict With India’s AI Governance Framework

Another major criticism from tech companies is that the proposal contradicts India’s recently released AI governance framework. Published on November 5 by a sub-committee led by the Principal Scientific Advisor, the framework encourages a balanced approach that supports TDM while safeguarding copyright holders’ interests. Although not legally binding, it was expected to guide the DPIIT committee’s approach. Instead, the IT ministry has backed the DPIIT recommendation despite the contradiction.

A Growing Policy Divide

The debate underscores a widening rift between government bodies, industry stakeholders and global technology companies over how to regulate AI training. With lawsuits already underway internationally—such as those involving OpenAI and Perplexity AI—India’s eventual policy choices will have wide-ranging implications for innovation, compliance costs, and the future of AI development in the country.

As Nasscom’s full dissent enters the public record, the policy conversation is likely to intensify, with industry groups pushing for flexibility and the government focused on protecting copyright holders and strengthening India’s intellectual property regime.