Hybrid AI Licensing Proposal by DPIIT Committee

India has entered the worldwide copyright and synthetic intelligence (AI) debate with the discharge of a paper titled Working Paper on Generative AI and Copyright Half 1: One Nation One License One Fee, printed by a committee fashioned by the Division for Promotion of Trade and Inside Commerce (DPIIT).

The report proposes a hybrid licensing framework that offers builders automated entry to all lawfully accessed copyrighted content material for coaching. In return, creators obtain statutory royalties routed via a single nationwide royalty mechanism, framed within the paper as a ‘One Nation One License One Fee’ method to AI coaching rights.

For context, the proposal marks India’s first formal try to resolve a fast-growing authorized battle between AI innovation and copyright safety. And it arrives at a second when lawsuits internationally are difficult builders who practice AI techniques on copyrighted materials with out permission, with India additionally going through comparable questions via ANI’s lawsuit in opposition to OpenAI within the Delhi Excessive Court docket (HC).

Notably, the advice alerts a deliberate coverage shift. The federal government desires to forestall licensing obstacles from slowing AI growth, whereas additionally making certain that creators acquire monetary profit from business AI techniques skilled on their work.

Furthermore, by changing fragmented licensing with a centralised cost and royalty distribution system, the framework makes an attempt to standardise compensation throughout sectors. By selecting this path, India positions itself between the extra permissive frameworks in Japan and Singapore, and the stricter compliance-focused guidelines rising within the European Union (EU). Elsewhere, the Indian authorities can also be constructing home AI functionality as a part of the IndiaAI Mission.

How the Hybrid Mannequin Works?

On the centre of the proposal is a compulsory blanket licence. As soon as builders receive lawful entry to copyrighted materials, they’ll use it for AI coaching with out negotiating particular person permissions or licensing contracts. In impact, creators can’t block the usage of their works for mannequin coaching. As an alternative, the system compensates them via statutory royalties.

To implement this construction, the committee proposes establishing a centralised entity known as the Copyright Royalties Collective for AI Coaching (CRCAT). This entity would gather funds from builders and distribute royalties to creators. Importantly, builders would contribute primarily based on pre-defined royalty formulation and income thresholds.

Moreover, creators should register their works to obtain payouts. Unregistered works can nonetheless be used for coaching, however creators wouldn’t be eligible to obtain compensation. Notably, registration doesn’t affect dataset entry.

The lawful entry requirement kinds the compliance boundary. Builders should buy, license, subscribe to, or in any other case legally entry content material earlier than utilizing it. Notably, the committee explicitly separates entry rights from copyright permission. As soon as lawful entry exists, no extra approvals are required.

Importantly, this proposal applies in a forward-looking method. Due to this fact, ongoing scraping disputes and previous unauthorised makes use of stay topic to present legislation and lively litigation.

Why the Committee Rejected Different Fashions?

The committee reviewed a number of regulatory approaches earlier than proposing the hybrid construction. First, it rejected a blanket textual content and knowledge mining (TDM) exception, as that mannequin permits builders to make use of copyrighted content material with out compensating creators. In line with the committee, such a system weakens inventive incentives and will increase the danger of AI outputs competing instantly with authentic works.

Subsequent, it examined the European-style TDM exception with opt-out rights. Nonetheless, the committee concluded that this mannequin was not workable as rights holders can’t meaningfully implement opt-outs until builders disclose detailed coaching datasets. The committee additionally warned that necessary dataset transparency might expose proprietary knowledge and impose heavy compliance burdens, particularly on smaller corporations.

The committee additionally dismissed voluntary licensing. In its view, negotiating with thousands and thousands of rights holders is unworkable at scale. Equally, it rejected prolonged collective licensing as India doesn’t but have a mature or unified licensing ecosystem. Many casual and neighborhood creators would stay exterior such a system, creating structural inequity.

Moreover, the committee evaluated conventional statutory licensing, which already exists in India’s broadcasting framework. Nonetheless, it decided that figuring out and compensating thousands and thousands of creators can be impractical and not using a centralised mechanism, and would probably recreate the identical transaction-cost obstacles that the coverage goals to take away.

Commercials

By means of elimination, the committee has concluded that the hybrid mannequin affords probably the most predictable construction for entry, compensation, and long-term implementation.

Authorities and Trade Positions

The Ministry of Electronics and Info Know-how (MeitY) helps the hybrid mannequin. In line with the ministry, builders want broad and consultant datasets to enhance mannequin efficiency and scale back bias. It additionally argues that creators ought to obtain compensation as AI-generated outputs more and more replicate identifiable creative types and artistic signatures. 

To stability innovation with equity, MeitY recommends safeguards comparable to triggering royalties solely after a mannequin or developer crosses an outlined income threshold. Moreover, it expects CRCAT to keep up clear reporting, predictable royalty formulation, and a structured course of for dispute decision.

Nonetheless, business teams take a distinct place. For context, Nasscom opposes the hybrid mannequin and argues that it provides administrative load which will sluggish innovation and disproportionately have an effect on smaller corporations. In its view, builders would wish new techniques to doc lawful entry, calculate royalties, and handle compliance, which might imply pointless value and operational friction. 

As an alternative, Nasscom proposes a authorized textual content and knowledge mining exception for each business and non-commercial use. Below this method, rights holders would use machine readable alerts to choose out of public datasets, whereas contracts would govern coaching on non-public content material. Nasscom argues that this mannequin protects rights with out creating a brand new licensing forms or compliance burden.

Why This Issues:

This proposal marks the start of a broader reform of India’s copyright framework within the context of AI. To elucidate, the subsequent part will handle unresolved questions round authorship of AI-generated content material, possession, ethical rights, and legal responsibility when outputs infringe copyright or trigger hurt. The federal government additionally plans to open the paper for public session earlier than drafting amendments to the Copyright Act.

Furthermore, the hybrid mannequin provides builders predictable rights to make use of coaching knowledge whereas aiming to make sure that creators obtain compensation. Nonetheless, it additionally creates a brand new licensing authority and introduces compliance obligations which will enhance operational complexity. Smaller corporations, analysis establishments, and open-source communities could really feel this influence first. Moreover, the formal dissent from business alerts that implementation is probably not easy and will face robust resistance throughout session.

If adopted, it could additionally form how Indian builders entry datasets and the way monetary worth is distributed throughout the AI ecosystem. Lastly, it is going to check whether or not India leans towards innovation pace, creator safety, or regulatory management as AI growth scales.

What Stays Unclear?

There are nonetheless questions within the air that want answering on AI and copyright. A few of them are as follows:

• How will builders show lawful entry, and what proof will regulators require?
• Will the framework require disclosure of coaching datasets, or will monetary and royalty reporting be ample?
• Will royalty obligations apply solely throughout preliminary coaching, or additionally throughout retraining, fine-tuning, and model updates?
• Will the framework create diminished compliance tiers for startups, analysis establishments, and open-source initiatives?
• How will the system handle coaching knowledge sourced internationally, the place copyright guidelines differ throughout jurisdictions?

Additionally Learn: