In a global first, India suggests AI firms like OpenAI and Google pay creators through a blanket license system — triggering debate over innovation, copyright, and fairness.
A Bold Move to Reshape AI Training Economics
India has introduced a sweeping proposal that would require AI companies to pay royalties when training their models on copyrighted content. The plan, unveiled by the Department for Promotion of Industry and Internal Trade, outlines a mandatory blanket licensing regime that could dramatically reshape the AI landscape, especially in one of the world’s fastest-growing digital markets.
- AI developers would be granted automatic access to copyrighted works.
- In return, they must pay into a centralized royalty-collection body.
- Funds would then be distributed to rights holders, including musicians, writers, and artists.
This approach marks a departure from the legal uncertainty surrounding AI training in the U.S. and EU, where the issue remains entangled in lawsuits and debates over fair use and text-and-data mining (TDM) exemptions.
A System Designed to Protect Creators — and Speed Compliance
The proposal, developed by an eight-member government committee, argues that a blanket license would create a “least burdensome” compliance route for AI companies while ensuring fair compensation for content creators.
- It eliminates the need for individual negotiations with rights holders.
- Both registered and unregistered creators would benefit.
- The new collecting society would function as a “single window”, streamlining royalty payments.
The committee frames this model as a “balanced framework” that ensures creators benefit from the value AI firms derive from Indian users and Indian content — a reference to OpenAI CEO Sam Altman’s remark that India is OpenAI’s second-largest market, potentially on track to become the largest.
Industry Pushback: Concerns Over Innovation and Implementation
Despite its creator-friendly design, the proposal has drawn sharp criticism from the tech industry. Major trade associations argue the mandatory license could stifle AI innovation.
- Nasscom, representing Google, Microsoft, and others, urged a broad TDM exception for AI developers using lawfully accessed material.
- The Business Software Alliance (BSA) — representing Adobe, AWS, and Microsoft — warned that relying solely on licensing would limit training data diversity, hurting model quality.
They advocate for either:
- A TDM exception, similar to Japan’s and Singapore’s laws.
- Or an opt-out system, allowing rights holders to remove their content from training datasets.
Both groups argue that mandatory payments for all content, regardless of creator intent or usage, create a disincentive to innovation.
Legal and Global Context: India vs. the West
India’s proposal lands as global legal battles heat up over the use of copyrighted content in AI training.
- In India, news agency ANI sued OpenAI, arguing that its content was used without consent — prompting the Delhi High Court to assess whether AI training is a form of reproduction or protected “fair dealing.”
- In the U.S. and Europe, lawsuits from authors, artists, and media groups question whether training on publicly accessible but copyrighted content violates IP laws.
Unlike the uncertain, court-driven approach in the West, India’s framework is proactive — offering a clear regulatory structure that ensures compensation without lengthy litigation.
A Hybrid Approach — But No Opt-Out
The Indian committee rejected both a TDM exception and an opt-out model, claiming they either:
- Weaken copyright protection, or
- Are too difficult to enforce at scale.
Instead, it recommends a hybrid system where:
- AI companies can train freely on lawfully available copyrighted material.
- But must pay royalties via a statutory license — no exceptions.
This makes India the first major market to propose a system that assumes AI companies owe creators by default — rather than letting them claim “fair use” without compensation.
What’s Next: Public Consultation and Finalization
The Indian government has opened the proposal to 30 days of public feedback. After this period, the committee will finalize its recommendations, and the framework may move to official policy or legislation.
- If adopted, this could set a global precedent, prompting other countries to consider similar royalty-based licensing models.
- It would also force AI firms to rethink their data strategies, especially in creator-rich, fast-growth markets like India.
So far, Google and OpenAI have declined to comment on the proposal.
