The federal government has proposed that know-how firms utilizing content material creators’ information to coach their synthetic intelligence (AI) and huge language fashions (LLMs) pay royalties to those creators at a charge set by the federal government or a court docket.
The brand new framework, proposed by the Division for Promotion of Trade and Inside Commerce (DPIIT), beneath the Ministry of Commerce and Trade, observes that fee for this blanket licence ought to be collected by a unified trade physique, which might then be accountable for distributing the proceeds amongst content material creators.
DPIIT launched these pointers as a part of a working paper late Monday night time, giving stakeholders 30 days to submit their views. In its draft report, the DPIIT rejected a voluntary licensing mechanism, arguing that such a framework would impose a “important compliance burden on these searching for licences, as a result of complexity of negotiations and the related uncertainty”.
A government- or court-fixed charge beneath a statutory licensing mechanism would, as an alternative, reduce prices and create a “predictable atmosphere for licensees of works”. “Whereas this mannequin takes away the ability of copyright house owners to refuse licensing or negotiate a payment, it ensures them truthful compensation,” the DPIIT report stated.
The statutory licensing framework may also make sure that AI fashions are freed from bias or hallucinations, as the utmost quantity of content material could be out there at affordable charges to all firms coaching AI fashions and LLMs, the working paper urged.
The framework also needs to enable the events — each firms that use content material to coach AI and LLMs, and content material creators — to problem or search a evaluation of the pre-determined charge earlier than a judicial discussion board.
Specialists, nevertheless, imagine the working paper’s premise — that truthful compensation for content material creators will be virtually and precisely administered by a compulsory collective mechanism for AI coaching — is flawed.
“The very mechanism supposed to simplify licensing and scale back transaction prices for AI builders creates an intractable ‘black field’ drawback within the royalty distribution section. Modern generative AI techniques course of large, various datasets in iterative and non-deterministic ways in which inherently don’t protect work-level traceability or contribution,” stated Jameela Sahiba, affiliate director at know-how coverage advocacy physique The Dialogue.
Whereas the hybrid mannequin could initially enhance firms’ compliance prices, it might result in higher record-keeping, cleaner licensed datasets, and standardised methods to trace information sources, stated Neelima Vobugari, cofounder and chief working officer at AiEnsured.
Different consultants imagine that, although the intent behind the framework is optimistic, it might create important points if applied in its present type. “The regulatory load will probably be huge, and anticipating a central company to handle filings, confirm claims, set royalties, and act because the intermediary between creators and AI firms feels unmanageable. Nearly each resolution will nonetheless be open to judicial evaluation, so who will realistically take that danger?” stated Pawan Prabhat, cofounder of Shorthills AI.
For full paper, https://www.dpiit.gov.in/static/uploads/2025/12/ff266bbeed10c48e3479c941484f3525.pdf
Enterprise Customary
Leave a Reply