India proposes a system that would make it mandatory to pay royalties for copyrighted content that can be used for AI learning

A proposed system in India would allow companies to charge fees for training AI on copyrighted content, a move that could significantly change how AI companies like OpenAI and Google operate in one of the world's most important and fastest-growing markets.

Press Release:Press Information Bureau

https://techcrunch.com/2025/12/09/india-proposes-charging-openai-google-for-training-ai-on-copyrighted-content/

On Tuesday, December 9, 2025 (local time), the Department for Promotion of Industry and Internal Trade (DPIIT) of India announced a framework to allow AI companies access to all copyrighted material for training their models. This will require AI companies to pay royalties to a new collection agency made up of rights-holding organizations. The royalties paid by AI companies will be shared with content creators.

If this proposal is implemented, DPIIT argues that it would reduce compliance costs for AI companies and ensure that content creators, who are copyright holders, are properly compensated if their content is used to train commercial models.

The use of copyrighted content for AI training has become a global issue. Recently, OpenAI's Sora 2 video generation AI was criticized for being able to generate Japanese content such as Pokémon and Mario.

OpenAI's video generation AI, Sora 2, is generating a large number of users who create videos of Pokemon, Mario, etc., and it is reported that OpenAI is preparing an option to exclude copyrighted content - GIGAZINE

In response, OpenAI CEO Sam Altman promised to improve the situation , and the Japanese government's request to OpenAI to 'avoid any acts that constitute copyright infringement' became a hot topic overseas.

The Japanese government has called Japanese intellectual property such as anime and games 'irreplaceable treasures' and officially requested OpenAI not to infringe on copyrights, which has become a hot topic overseas - GIGAZINE

OpenAI, a frontrunner in AI development, is also being sued for allegedly using copyrighted books as learning materials without permission.

Regarding India's policy, technology media TechCrunch pointed out, 'Unlike the US and EU, where policymakers are debating transparency mandates and the boundaries of fair use , India is proposing one of the most interventionist approaches yet: giving AI companies access to copyrighted content in exchange for mandatory content royalties.'

OpenAI CEO Altman said that India is the company's second largest market after the United States and 'has every chance of becoming the largest market.' The committee promoting the new system argued that AI companies earn significant revenue from content created by Indian creators, and that 'a portion of the revenue AI companies earn should be returned to creators.' Therefore, they stated that a 'balanced framework' like the one proposed this time, which guarantees compensation, is necessary.

However, NASSCOM , an industry association representing India's major IT companies, expressed its opposition to the new system, arguing that India should introduce broad exemptions for text and data mining. NASSCOM warned that a mandatory licensing system could stifle innovation, and argued that companies should be allowed to opt out of using all training data, rather than being required to pay for it.

The Business Software Alliance, a nonprofit organization working with the business software industry, urged avoidance of a license-based system, arguing that 'relying solely on direct or statutory licenses for AI training data is unrealistic and may not produce the best results.'

The Indian government is inviting public comments on the proposal, inviting companies and other stakeholders to submit their views within 30 days. A committee will then consider the feedback and make final recommendations before the government adopts the framework.