AI Laws in India:
The landscape of Indian copyright law is currently undergoing its most significant transformation since the 2012 amendments. Driven by the ANI Media v. OpenAI litigation and the broader "AI for All" mission, the Indian government is shifting from a wait-and-see approach to active legislative intervention.
Below is a breakdown of the core shifts and current legal status based on the May 2025 expert panel's progress and recent judicial signals.
1. The 2025 Reform Blueprint: "One Nation, One License."
The Department for Promotion of Industry and Internal Trade (DPIIT) recently shifted from asserting that current laws are "sufficient" to proposing a radical new Hybrid Licensing Model in late 2025.
- Mandatory Statutory Licensing: The panel has proposed a "One Nation, One License" framework. Instead of AI companies negotiating with every individual publisher (which is seen as logistically impossible), they would receive a blanket license to use lawfully accessed content for training.
- Revenue-Share Royalties: Unlike traditional copyright fees, the proposal suggests royalties based on a percentage of the global gross revenue of the commercialized AI system.
- The CRCAT: A new body, the Copyright Royalties Collective for AI Training (CRCAT), is proposed to manage these collections and distribute them to creators.
2. Judicial Watershed: ANI Media (P) Ltd. v. OpenAI
The Delhi High Court's handling of the ANI case (2024–2025) has become the primary laboratory for these issues:
- The "Opt-Out" Reality: In October 2024, OpenAI blocklisted ANI's domain to prevent further scraping. However, the court is currently debating whether past training on that data constitutes a finished "infringement" that requires compensation.
- Fair Dealing Under Fire: While OpenAI argues that training is "transformative" and falls under Section 52 (Fair Dealing), the DPIIT and the court’s amici curiae have signalled that commercial-scale data mining for profit does not fit the traditional "private or research" exception.
3. The Authorship Gap (Human vs. AI):
Indian law remains strictly anthropocentric. As of early 2026:
- AI-Generated: Works created entirely by AI without human intervention fall into a "legal vacuum" and are currently denied copyright protection.
- AI-Assisted: The panel is moving toward a model similar to the UK’s Section 9(3), where the "author" is the person who made the "arrangements necessary" for the creation of the work.
- Rule 83A Amendment (June 2025): New procedural rules now mandate exclusively digital payment systems for copyright licensing, setting the stage for the automated royalty distributions required by AI training models.
Comparative Status Summary (2025-2026):
| Authorship | AI Training | Transparency | |
|---|---|---|---|
| Strictly Human (Moving to 'Arranger' model) | Proposed Statutory License (Paid) | Mandatory Dataset Disclosure | |
| Strictly Human | Fair Use (Subject to Litigation) | Voluntary / Proposed Bills | |
| Human-centric (AI Act focus) | TDM Exception (with Opt-out) | Mandatory (EU AI Act) |
The proposed Chapter XII-A of the Copyright Act, 1957, represents a foundational shift from "protection-by-permission" to a "remuneration-based" ecosystem. For individual digital creators—graphic designers, writers, coders, and musicians—this transition will fundamentally change how they control, monetize, and protect their work.
1. The End of "Opt-Out" for Training:
The most significant shift in the 2025–26 proposal is the Mandatory Blanket License.
- The Change: Previously, you could theoretically sue a company for "scraping" your portfolio to train an AI. Under the new Chapter XII-A, AI developers will have a statutory right to use any "lawfully accessed" content (anything public on the web) for training.
- The Impact: You can no longer prevent your style or voice from being "learned" by an AI if your work is public. However, this is balanced by a new right to get paid.
2. Transition to "Passive Income" via CRCAT:
Instead of chasing individual tech giants for licensing fees, the government is introducing the Copyright Royalties Collective for AI Training (CRCAT).
- Micro-Royalties: AI companies will pay a percentage of their commercial revenue into this central pool.
- Automated Distribution: As a creator, you will likely need to register your "digital fingerprint" or portfolio with a recognized Copyright Society. When an AI model is commercialized, royalties are distributed to registered creators based on the "contribution" of their data class.
- Rule 83A Integration: All these payments must be digital and transparent, reducing the "opaque" delays typical of older royalty systems in India.
3. The "Arranger" Model of Authorship:
The panel is moving toward the UK-style Section 9(3) logic to solve the "Who owns the output?" dilemma.
- AI-Assisted vs. AI-Generated: If you use AI as a tool (like a smart brush in Photoshop) and provide "significant creative arrangement," you remain the author.
- If you simply click "generate" with a one-word prompt, the work may enter the public domain or be attributed to the person who "made the arrangements" (potentially the developer or the enterprise user).
- Strategy for Creators: To ensure you own the copyright, you must document your iterative process (the specific prompts, edits, and "human-in-the-loop" decisions) to prove you are the "arranger."
4. New Transparency & Labelling Mandates:
Under the February 2026 IT Rules (which work in tandem with the Copyright Act), the burden of "authenticity" has shifted.
- Mandatory Labelling: If you use GenAI to create commercial content, you must clearly label it as "Synthetically Generated Information" (SGI).
- Digital Fingerprinting: Platforms are now required to embed metadata that traces the work back to its source. For you, this means better protection against "deepfake" clones of your work, as the origin of the content becomes legally traceable.
5. Reversal of the "Burden of Proof."
In a radical move, the panel has suggested that in infringement cases, the AI Developer must prove they didn't use your work if the output is "substantially similar."
- The Win: For an individual creator, this is massive. You no longer need to reverse-engineer a black-box AI to prove they scraped your art; if the AI's output looks like your unique style and you are registered, the company must prove their "innocence."
Summary Table for Digital Creators (2026):
| Control: You decide who trains on your art. | Compensated Access: You can't stop the training, but you must be paid. |
| Monetization: Negotiate individual licenses. | Collective Royalty: Automated payouts via CRCAT. |
| Authorship: Only "Human" works protected. | Arranger Rights: Human "arrangements" of AI work are protected. |
| Infringement: You must prove the AI scraped you. | Presumed Use: AI firm must prove they didn't use your work. |
Navigating the New Era with JTS Lex:
As these legislative interventions under the 2025 Reform Blueprint and Rule 83A continue to mature, the role of specialized legal counsel becomes indispensable. JTS Lex remains at the forefront of this transformation, providing comprehensive overviews and strategic guidance for both tech innovators and individual creators. By bridging the gap between traditional copyright principles and the complexities of generative AI, JTS Lex ensures that clients can navigate the "One Nation, One License" framework with clarity, ensuring their digital assets are both legally compliant and optimally monetized in this new Indian "AI for All" era.
