All Things Policy One License, Many Problems
9 snips
Feb 4, 2026 Bharath Reddy, researcher on AI policy and copyright at the Takshashila Institution, breaks down India's 'One-Nation, One-License' proposal. He discusses how mandatory blanket licensing could reshape AI training, the legal and practical challenges of attributing model outputs, and the economic risks of fees and disclosure rules. He contrasts alternative approaches like opt-outs and collective deals.
AI Snips
Chapters
Transcript
Episode notes
Transformative Defense Shapes US Copyright Rulings
- US courts often treat generative AI as transformative and not substituting original markets, reducing clear-cut copyright infringement claims.
- This legal view makes mandatory compensation schemes legally questionable and unsettled.
AI Can Divert Traffic From Publishers
- Generative AI can divert user traffic away from original publishers by answering queries directly.
- That traffic loss threatens publishers' ad revenue and motivates calls for creator compensation.
Global Approaches Vary Widely
- Different jurisdictions adopt varied approaches like machine-readable opt-outs and broad text-and-data-mining exemptions.
- Collective licensing is another model where publishers pool rights and share proceeds with AI developers.