Govt Working Paper Proposes Royalty Model for AI Access to Online Content
A government working paper has proposed a fundamental shift in how India regulates the use of online content for training large language models. The draft suggests that AI developers should have default access to freely available web content, with no opt-out mechanism for publishers, while a royalty-collecting non-profit would distribute payments to rightsholders.
Key Recommendations of the Working Paper
The report, prepared by a committee under the Department for Promotion of Industry and Internal Trade, argues that publishers should not be allowed to block AI systems from mining publicly accessible content. Instead, a copyright-society-like entity would collect royalties on behalf of both members and non-members, ensuring compensation without requiring content owners to negotiate individually.
Nasscom’s Dissent and Industry Concerns
Nasscom opposed the recommendation, stating that mandatory royalty payments would function as a “tax on innovation.” It argued for permitting unrestricted mining of non-paywalled content and insisted that publishers of both open and restricted websites should have the ability to reserve their material from AI training. The committee rejected these objections, citing enforcement challenges for small creators.
Impact on Copyright Disputes and AI Governance
The proposal comes amid ongoing legal disputes, including a lawsuit by the Digital News Publishers Association against an AI developer for alleged copyright infringement. The paper suggests that relying on lengthy litigation is impractical and that a statutory model could resolve questions around data access by legitimising training use in exchange for fees, similar to compulsory licensing in radio broadcasting.
Exam Oriented Facts
- The DPIIT working paper proposes default access for AI systems to freely available online content.
- A non-profit body would collect and distribute royalties to all content creators.
- Nasscom dissented, calling compulsory royalties a barrier to innovation.
- The model draws on India’s compulsory licensing framework used for radio music rights.
Royalty Distribution and Potential Pushback
Payments collected by the proposed non-profit would be allocated using metrics such as web traffic and publisher reputation. Both AI developers and content creators are expected to debate the framework: one side wary of increased operational costs, the other fearing undervaluation of their contributions. Any disputes under the new regime would be subject to judicial appeal, ensuring oversight as India shapes its AI-copyright policy.