DPIIT Proposes 'Mandatory Blanket License' Hybrid Model to Balance AI Innovation and Creator Rights
SDG 9: Industry, Innovation, and Infrastructure | SDG 16: Peace, Justice and Strong Institutions
Department for Promotion of Industry and Internal Trade (DPIIT) | Ministry of Commerce & Industry | Ministry of Electronics and Information Technology (MeitY)
The DPIIT Working Paper (Part 1), titled “One Nation, One License, One Payment: Balancing AI Innovation and Copyright”, examines a foundational challenge in generative AI—the use of vast amounts of copyrighted material to train AI systems. Authored by the newly constituted Committee on Generative Artificial Intelligence and Copyright, the paper evaluates whether India’s existing copyright framework is adequate to address the legal and policy questions raised by this emerging technology.
Core Legal Conflict (Input Side):
The central issue is whether the act of Text and Data Mining (TDM)—copying and processing works for AI training—constitutes infringement under the Indian Copyright Act, as there is no specific exception for it. While the tech industry advocates for a blanket “Fair Use” exception, the content industry demands compensation, citing major legal ambiguity (e.g., the pending ANI Media Pvt. Ltd. v. Open AI Inc. case).
Proposed Hybrid Solution (Majority View):
The Committee rejected both a full ‘Fair Use’ exception (which undermines creator compensation) and a pure voluntary licensing model (which creates high transaction costs and risks biased models). Instead, it proposes a Hybrid Model:
Mandatory Blanket License: AI Developers gain a statutory right to use all lawfully accessed copyrighted content for training their systems without prior permission or negotiation. This ensures access to broad, high-quality data to mitigate AI bias and hallucinations.
Statutory Remuneration: Copyright holders receive compensation via an unwaivable statutory remuneration right.
CRCAT Collection Mechanism: Royalties are collected and administered by a new, centralized non-profit umbrella entity, the Copyright Royalties Collective for AI Training (CRCAT).
Revenue-Based Rates: Royalties are calculated as a fixed percentage of the AI System’s gross global revenue (excluding taxes), set by a Government-Appointed Committee. This revenue-share model ensures no upfront cost for AI startups.
Distribution & Inclusivity: CRCAT distributes royalties to its member Copyright Societies (CS) and Collective Management Organisations (CMOs), and crucially, will hold royalties for the currently unorganized sectors for three years to allow them to form CMOs and claim their share.
What is the CRCAT Hybrid Model’s Core Mechanism? The Hybrid Model works as an automatic, mandatory licensing and payment system. An AI Developer who lawfully accesses works trains their model, and upon Commercialization, submits an AI Training Data Disclosure Form (summarizing data categories/sources) to CRCAT, and pays a revenue percentage set by a Government Rate Setting Committee16. This shifts the regulatory focus from complex pre-use permission to post-commercialization compensation and transparency.
Policy Relevance
This DPIIT working paper is crucial for India’s economic future, balancing two pillars of growth:
Creative Economy Protection: The creative and media sector contributes 0.73% of India’s GDP, and the informal music industry employs over 1.4 crore people. The Hybrid Model explicitly prevents the substitution effect where creators are driven out of the market by unpaid AI, ensuring that creators are fairly compensated and retain the incentive to produce the high-quality, diverse content needed to train future AI systems.
AI Innovation & Viability: The framework gives India a competitive edge by avoiding the legal uncertainty of the US and the opt-out mechanism of the EU (which risk limited, biased datasets). By mandating no upfront fees and reducing transaction costs, the model directly supports India’s status as the second-largest market for platforms like OpenAI and promotes the government’s IndiaAI Mission to create indigenous foundational models (LLMs/SLMs).
Follow the full report here: ONE NATION ONE LICENSE ONE PAYMENT: Balancing AI Innovation and Copyright