On December 9, the Department for Promotion of Industry and Internal Trade (“DPIIT”) released a working paper on Generative AI and Copyright- One Nation One License One Payment Balancing AI Innovation and Copyright (“Working Paper”).
SFLC.in brief concerns are along the following lines:
- In the mandatory blanket licensing model, creators lose autonomy over their content and are forced to take part in AI training without an opt-out or opt-in option. There are also no safeguards for personality rights.
- The mandatory blanket license model ignores Free and Open Source Software (FOSS) and open licenses, which are built on “voluntary sharing” and “share-alike” principles. Mandatory licensing allows proprietary AI systems to absorb open-source work without respecting downstream license obligations. It also forces creators and developers who intentionally share their work for public benefit to accept compensation and undermine FOSS ideologies.
- The proposed flat, revenue-linked remuneration model does not offer fair or meaningful compensation to creators. Creators are asked to surrender their works upfront while compensation remains speculative, depending on the AI company’s revenue and are likely negligible.
- It narrows the scope of existing copyright infringement remedies (damages, injunctions) by limiting it to unlawful access and royalty disputes.
- The Working Paper proposes solutions for AI training without first defining what constitutes copyright infringement in generative AI models.
SFLC.in proposes the following solutions:
- Adopt an opt-in license model for AI training, so that creators have autonomy to choose whether their content is available for AI training, and respect the existence of different licensing models such as open-source licenses.
- Introduce upfront payment for training data, plus revenue-based royalties, once the AI company generates revenue. This prevents creators from bearing the risk of AI experimentation.
- Limit the mandatory blanket license model to public-interest sectors like education, healthcare, and non-commercial research, and it must be narrowly tailored, purpose-bound, and justified by demonstrable public interest.
Read the detailed comments here:
