India's traditional textile motifs — Banarasi, Kanchipuram, Sanganeri, Bagru — are entering generative AI systems that can reproduce them infinitely. The core tension is not whether AI can replicate a weaver's flower pattern. It can. The question is who captures the economic value when it does, and current Indian law offers no answer. The extraction mechanics are already visible at the artisan level. Weavers in Sualkuchi receive a one-time fee of roughly 2,000 rupees ($21) per design with no royalties — a flat buyout of generational knowledge. Computer scientist Bitupan Deka built a system to generate infinite new designs from existing patterns, framing it as a tool for weavers. But the patent is being filed under his university, and the planned monetisation model is a commercial design service sold to the textile industry, not to artisans. Tata Consultancy Services' Bridgital Loom, showcased at the India AI Impact Summit in February, converts textile designs into loom-ready instructions using LED guidance. TCS frames this as a bridge technology that lets new weavers skip the 15-year mastery curve. The flip side: the weaver follows machine instructions, surrendering creative agency and pattern memory to the system. Technology compresses the master-apprentice hierarchy into a software layer owned by TCS. The Indian government's digitisation efforts — the Bharatiya Vastra Shilpa Kosh portal, the $1.4 million Centre of Excellence for Handloom Technology at IIT Delhi, plans to sensor-fit thousands of handlooms — create the very datasets that AI systems train on. Each act of preservation simultaneously makes extraction easier. Maharashtra's Vingatha system and proposed QR-coded provenance tags offer traceability, but traceability is not compensation. The legal gap is structural. India's December 2025 proposed blanket licence would let AI developers train on copyrighted material and pay government-set royalties. But textile motifs are communal property with no single traceable author. IP law requires an identifiable rights holder. A motif inherited across generations through oral tradition has none. The artisan whose name never appeared on a label will not appear on the royalty ledger either. Third-generation Kanchipuram weaver Ananthalakshmi Vadivel insists authentic handcraft will always command a premium from discerning buyers. Bagru designer Dinesh Chhipa accepts AI filling market gaps but draws the line at borrowing craft and regional names. Both positions assume a stable luxury niche for authentic work. That assumption may hold for the top tier, but the mass market — where most artisans actually earn — is precisely where AI-generated designs displace human creators at zero marginal cost. The emerging pattern is familiar from other extraction cycles: communal knowledge digitised into trainable datasets, IP frameworks that cannot recognise collective ownership, technology platforms capturing the commercial upside, and artisans offered visibility tools instead of revenue streams. The saree may tell you where it came from, but it cannot send money back.
