New Delhi (ABC Live): As India accelerates its push toward sovereign artificial intelligence, the legal framework governing how machines consume human creativity is now facing its most consequential test yet. At the heart of this transformation lies the DPIIT AI Copyright Plan, unveiled through the Working Paper on Generative AI and Copyright (Part I, December 2025) by the Department for Promotion of Industry and Internal Trade (DPIIT).

Notably, this proposal does not emerge in isolation. Instead, it coincides with three parallel national strategies that together define India’s AI ambition:

• Data Sovereignty as National Power – explained in detail in ABC Live’s analysis on India’s data-first AI strategy
  👉 https://abclive.in/2025/09/08/explained-india-data-sovereignty-before-ai-development/
• Chip Sovereignty through SEMICON India 2025 – India’s semiconductor and AI hardware backbone
  👉 https://abclive.in/2025/09/05/semicon-india-2025/
• Hyperscale AI Infrastructure via Google’s Visakhapatnam Data Centre Hub – turning India into a live global AI processing base
  👉 https://abclive.in/2025/10/15/google-data-centre-visakhapatnam/

Taken together, these three layers form India’s emerging AI–Data–Compute Sovereignty Stack. Consequently, the DPIIT AI Copyright Plan now attempts to add a fourth layer:
sovereign control over how copyrighted human creativity is monetised in the age of generative AI.

However, while the ambition is historic, the constitutional, economic, and institutional consequences are equally far-reaching.

What the DPIIT AI Copyright Plan Proposes

At the outset, it is important to clarify that Part I of the DPIIT Working Paper addresses only the “input side” of AI—in other words, how generative AI models use copyrighted material for training. Within this limited yet critical scope, the plan proposes five structural changes.

1. Mandatory Blanket Licence for AI Training

To begin with, the DPIIT AI Copyright Plan proposes that all lawfully accessible copyrighted content—including books, news articles, music, films, art, and photographs—shall become automatically available for AI training.

More significantly, creators will not have the legal right to refuse the use of their works for training. As a result, exclusive property rights at the training stage would be statutorily extinguished.

2. Statutory Royalty Only After Commercialisation

Next, the plan shifts the economic burden to the output stage. Accordingly:

• AI developers will not pay anything at the training stage.
• Instead, royalty becomes payable only once an AI system is commercially deployed.

Furthermore, royalty rates will be:

• Fixed by a government-appointed committee, and
• Subject only to judicial review, not direct parliamentary pricing oversight.

Thus, the economic compact of the DPIIT AI Copyright Plan operates on a “train now, pay later” principle.

3. One Centralised National Royalty Collective

In addition, the plan creates a new umbrella institution—the Copyright Royalties Collective for AI Training (CRCAT). This body will:

• Collect royalties from AI developers,
• Distribute them through copyright societies and CMOs, and
• Allow even non-member creators to receive payments if they register their works.

As a consequence, the entire AI–copyright economy would flow through a single national revenue clearinghouse.

4. Rejection of Global TDM Exception Models

At the same time, the Committee explicitly rejects:

• Blanket Text and Data Mining (TDM) exceptions,
• EU-style opt-out frameworks, and
• Voluntary bilateral licensing systems.

Instead, it argues that such regimes either weaken creators’ bargaining power or disproportionately benefit large technology platforms.

5. Recorded Dissent from Nasscom

However, the policy consensus fractures at this point. India’s leading tech industry body, Nasscom, formally dissented. In contrast, Nasscom demanded:

• A commercial and non-commercial TDM exception,
• Machine-readable opt-outs for public content, and
• Contract-based reservations for private content.

Therefore, even within India’s policy establishment, the DPIIT AI Copyright Plan already reflects a deep split between creator-first and developer-first approaches.

Why the DPIIT AI Copyright Plan Is Being Framed as “Digital Sovereignty”

Today, India is no longer regulating only platforms or intermediaries. Instead, it is now regulating the core extraction layer of the AI economy: data and creativity itself.

On the one hand, India’s data sovereignty doctrine holds that Indian data must serve Indian public interest first.
On the other hand, SEMICON India 2025 seeks domestic control over the physical chips powering AI.
Meanwhile, the Visakhapatnam hyperscale data centre positions India as a global AI compute hub.

Consequently, the DPIIT AI Copyright Plan now functions as the final sovereignty layer:

If AI is built in India, on Indian compute, using Indian data, then Indian creators must receive sovereign compensation—by law.

Therefore, this is not merely copyright reform. Instead, it represents the economic constitution of India’s future AI state.

Constitutional Risks Inside the DPIIT AI Copyright Plan

Despite its economic ambition, the plan triggers three direct constitutional flashpoints.

1. Article 300A – Forced Taking of Intellectual Property

First and foremost, copyright is recognised by the Supreme Court as constitutional “property.”

Yet, the DPIIT AI Copyright Plan explicitly states that rightsholders will not have the option to withhold their works from AI training. As a result, exclusive property is converted into compulsory public-use property—not for public health or education, but for commercial AI products.

This creates a classic regulatory taking problem:

• Property is taken,
• Consent is eliminated,
• Compensation is deferred, and
• Pricing is executive-controlled.

Accordingly, without a strict test of necessity and proportionality, the DPIIT AI Copyright Plan remains highly vulnerable under Article 300A.

2. Article 14 – Excessive Delegation & Manifest Arbitrariness

Secondly, royalty rates are to be fixed by a government committee. However, the Working Paper provides:

• No statutory pricing formula,
• No minimum or maximum limits,
• No guaranteed creator representation, and
• No transparent pricing principles.

Consequently, absolute pricing power concentrates in the executive, inviting:

• Regulatory capture,
• Intense industry lobbying, and
• Marginalisation of regional and folk creators.

Under settled constitutional doctrine, price fixation without legislative safeguards is vulnerable to being struck down as arbitrary.

3. Article 19(1)(g) – Freedom of Trade of Creators

In today’s digital economy, creativity is not merely cultural—it is also commercial. Exclusive licensing, platform syndication, and digital publishing constitute livelihood.

However, the DPIIT AI Copyright Plan:

• Removes the right to refuse AI training,
• Converts bargaining markets into state-priced monopolies, and
• Forces creators into a single national royalty pool.

As a result, it directly burdens the freedom to practise one’s profession and trade under Article 19(1)(g).

How the DPIIT AI Copyright Plan Compares with Global Models

Thus, the DPIIT AI Copyright Plan represents a globally unique triple-control structure:

Forced access + State pricing + Centralised distribution

No other major AI jurisdiction currently combines all three.

Economic Fault Lines in the DPIIT AI Copyright Plan

Although the plan claims to protect “small creators,” it also carries serious political-economy risks.

Collective Capture

Historically, large royalty pools benefit:

• Major music labels,
• Big broadcasters, and
• Dominant copyright societies.

Meanwhile, informal artists, folk performers, and regional journalists often remain under-represented.

Registration Bias

Only creators who successfully register their works will be paid—even if:

• Their content is scraped,
• Their dialect lacks documentary infrastructure, or
• Their work circulates informally through social media.

Opaque Distribution

Finally, the Working Paper provides no mathematical royalty-sharing formula across languages, genres, platforms, or regions. Consequently, sovereign extraction may not automatically translate into equitable compensation.

Why the DPIIT AI Copyright Plan Matters for India’s Grand AI Strategy

At present, India is simultaneously:

• Asserting data sovereignty,
• Building fabs under SEMICON India,
• Hosting global AI infrastructure through Google and other hyperscalers, and
• Attempting to price creativity through national law.

Together, this represents a civilisational redesign of technological capitalism. However, constitutional shortcuts at this foundational stage could invite:

• Immediate High Court and Supreme Court litigation,
• Foreign investment arbitration risks,
• WIPO and TRIPS treaty challenges, and
• Long-term regulatory uncertainty for India’s AI economy.

A More Constitutionally Stable Path Forward

Nevertheless, India could still lead globally by recalibrating the DPIIT AI Copyright Plan through:

• Layered opt-outs instead of total compulsion,
• Independent royalty tribunals instead of executive pricing,
• Transparent training-data disclosure categories,
• Direct output-side liability for AI memorisation and plagiarism, and
• Minimum guaranteed payouts for informal creative sectors.

Only then can India credibly claim to have built a rights-respecting, innovation-ready, and constitutionally stable AI economy.

Conclusion: DPIIT AI Copyright Plan at a Crossroads

In conclusion, the DPIIT AI Copyright Plan is India’s first serious attempt to constitutionally price creativity inside the economics of generative artificial intelligence. Undoubtedly, its ambition mirrors the scale of India’s AI aspirations. Nevertheless, in its present form, the plan remains:

• Legally fragile,
• Constitutionally vulnerable,
• Economically centralising, and
• Politically contested.

Ultimately, whether the DPIIT AI Copyright Plan evolves into a new sovereign social contract for India’s AI economy—or collapses into a litigation-heavy regulatory misfire—will depend on how Parliament, courts, creators, and AI developers reshape it in the months ahead.

Verified References (Exact Links)

1. DPIIT – Working Paper on Generative AI & Copyright (Part I, December 2025)
   https://www.dpiit.gov.in/static/uploads/2025/12/ff266bbeed10c48e3479c941484f3525.pdf
2. ABC Live – India’s Data Sovereignty Before AI Development
   https://abclive.in/2025/09/08/explained-india-data-sovereignty-before-ai-development/
3. ABC Live – SEMICON India 2025
   https://abclive.in/2025/09/05/semicon-india-2025/
4. ABC Live – Google Data Centre Visakhapatnam
   https://abclive.in/2025/10/15/google-data-centre-visakhapatnam/
5. IndiaAI Mission – Official Portal
   https://indiaai.gov.in/
6. PIB – IndiaAI Mission Approval (₹10,300 Crore)
   https://www.pib.gov.in/PressReleasePage.aspx?PRID=2012375
7. EU DSM Directive (Copyright)
   https://eur-lex.europa.eu/eli/dir/2019/790/oj
8. Berne Convention
   https://www.wipo.int/treaties/en/ip/berne/
9. TRIPS Agreement – Article 9
   https://www.wto.org/english/docs_e/legal_e/27-trips_01_e.htm

How We Verified This Report

This analysis is based on the official DPIIT Working Paper on Generative AI and Copyright (Part I, December 2025), cross-referenced with authenticated IndiaAI Mission documentation, Press Information Bureau (PIB) releases, European Union copyright law under the DSM Directive, and ABC Live’s independent investigative reporting on India’s data sovereignty framework, semiconductor strategy under SEMICON India 2025, and hyperscale AI infrastructure developments, including Google’s Visakhapatnam data centre hub.

All legal interpretations are grounded in established constitutional doctrine, Supreme Court jurisprudence on intellectual property and property rights, and comparative international copyright regimes.