Indian Government proposes retroactive AI royalties for Google, OpenAI

India is considering a major, precedent-setting shift in AI regulation, with a government panel proposing that global companies like Google and OpenAI compensate content creators when their work is used to train artificial intelligence models. The recommendation, released in a report on December 8, could financially impact tech giants that have long relied on free access to publicly available data.

The working paper, titled “One Nation One Licence One Payment: Balancing AI Innovation and Copyright,” was published by a committee under the Department for Promotion of Industry and Internal Trade (DPIIT).

The new framework: mandatory blanket license

Under the proposed framework, AI firms would retain access to lawfully accessed Indian content for model training via a “mandatory blanket license” but would be required to pay royalties into a centralized fund representing copyright holders.

This approach, known as a Hybrid Model, aims to ensure creators receive a fair share of the revenue generated by AI tools built using their work. The fund would be managed by a government-designated non-profit, such as the proposed Copyright Royalties Collective for AI Training (CRCAT).

The retroactive payment clause

Crucially, the panel recommended that the royalty obligation should apply retroactively to commercially successful AI systems already trained on copyrighted content. This provision is intended as a corrective mechanism to ensure fairness for creators whose works have been used by high-revenue AI models developed over the past few years.

A divergence from global models

India’s plan diverges significantly from existing frameworks in the West:

• United States: AI companies largely argue that using public content falls under “fair use” and requires no compensation.
• European Union: Relies on an “opt-out” system, where creators must proactively block their work from being used.

The Indian panel found the EU’s opt-out approach ineffective, noting that it places an unfair burden on creators to track their content across massive, opaque datasets. Instead, India proposes a model where creators can claim compensation from a central pool if their work is utilized. The proposal is now open for a 30-day public and industry consultation.

Industry pushback and legal context

Reactions from the technology sector have been swift. Nasscom, the industry body representing major players including Google and Microsoft, strongly opposed the plan, characterizing the mandatory fees as a “tax or levy on innovation.” Conversely, the Motion Picture Association, which includes members like Netflix, urged that current copyright laws remain unchanged, favoring licensing deals over government mandates.

The proposal emerges amid intensifying legal disputes in India, notably the ongoing court case between OpenAI and Indian news agency ANI over alleged unauthorized use of copyrighted content.

Brief summary:

India is exploring a new AI regulation framework that would require global tech giants to pay royalties to content creators for AI training. By mandating a blanket license with a statutory right to remuneration—including a potentially massive retroactive payment clause—the plan aims to protect creators while balancing the needs of a thriving AI innovation ecosystem.

FAQs

Q1: What is the core of the DPIIT panel’s proposal?

It recommends a “mandatory blanket license” allowing AI developers to use all lawfully accessed copyrighted works for training, but requiring them to pay royalties into a centralized fund (the Hybrid Model).

Q2: What is the most significant financial detail in the proposal?

The proposal suggests the royalty obligation should apply retroactively to commercially successful AI systems. This means existing AI models (like ChatGPT or Gemini) could be subject to back-dated payments for content used years ago.

Q3: Which government body released this proposal?

The working paper was released by a committee under the Department for Promotion of Industry and Internal Trade (DPIIT).

Q4: Who will manage the royalty fund?

The fund would be managed by a government-designated non-profit, tentatively named the Copyright Royalties Collective for AI Training (CRCAT).

Q5: How is India’s model different from the EU’s?

The EU relies on an “opt-out” system (creators must ask to be excluded). India proposes a royalty pool where creators are compensated automatically via the collective, removing the burden of tracking their content across datasets.

Q6: What is the next step for this policy?

The proposal is open for a 30-day public and industry consultation. After this period, government officials will review feedback before deciding on implementation.