AI and the question of who owns India's textile heritage

On a shopping trip with his mother for gamusa — the traditional scarf of India's northeast — in Meghalaya state, Bitupan Deka noticed something odd: every shop was selling the same few motifs, repeating over and over.

Digging deeper, he learned that weavers in Sualkuchi, a district in neighbouring Assam, were paid a one-time fee of about 2,000 rupees (around $21) per design they created, rather than royalties. Cutting the number of designs was a cost-saving move by manufacturers.

The discovery pushed Deka, now an assistant professor of computer technology at MIT University in Shillong, to build a system that can generate endless new designs for such fabrics. He insists his work is not meant to replace traditional designers, but to give them an easy tool for creating new, more modern motifs so consumers have more choices. Once a patent — being filed through the university's research and development office — is approved, Deka plans to commercialise the system as a design service the textile industry can use.

In India's diverse textile heritage, patterns and motifs carry the memory of a place, the myths of its people, even the history of migration. Weavers create patterns from what surrounds them, then colour them with local materials, mostly on clay blocks and carved wooden frames. They do not merely copy a motif but inherit it: their hands hesitate exactly where their grandmother's did, and they understand why a particular flower must face left.

A computer has no such inheritance or hesitation. It learns by analysing images and repeating them in endless variations. Yet, as Deka's system shows, a computer can now draw and weave that same flower.

The Rekh Sanganeri dataset is a similar system developed for the hand block-printed Sanganeri fabric of Rajasthan in western India. It contains high-resolution photographs of printed patterns, annotated in detail down to ink smudges and slight misalignments — imperfections that are the signature of handmade goods. Its developer, Reddy Sai Nivas C, insists the dataset should be used as research and artistic material, not to replace artisans.

But that very disclaimer contains a paradox: tools created to preserve a craft are making its visual language easier to reproduce digitally, without the craftsman.

Dinesh Chhipa comes from one of the oldest Chhipa families — a caste of artisans and traders famed for dyeing and block printing. He is a Bagru designer, practising the block-printing technique of the town of Bagru in Rajasthan, and a textile entrepreneur. Traditional craft, he says, produces slowly; however much someone pays, artisans cannot make large volumes, so he does not blame those who use digital technology and AI to fill the gap. Still, he objects to borrowing the names of the craft and the region, such as Bagru or Jaipuri, because those designs and motifs were born from the culture and needs of the community that owns them.

Who owns culture once it becomes data?

India has been digitising its textile memory for years, from the historical collections of the Botanical Survey of India to the Bharatiya Vastra Shilpa Kosh — a national portal built by the National Institute of Fashion Technology and the Centre for Development of Advanced Computing (C-DAC). Undocumented, these fragile traditions and knowledge could vanish. But digitisation also means that repository can be used by an AI system to reproduce crafts at scale.

At the India AI Impact Summit in New Delhi in February this year, software company Tata Consultancy Services (TCS) presented Bridgital Loom, an AI system that turns textile designs into instructions ready for the loom. LED lights on the loom guide the weaver through each threading step of intricate Banarasi and Kanchipuram motifs — named for the cities where those crafts took shape. TCS says the technology helps weavers execute complex designs faster and with fewer errors, and is seen as a bridge between artisans and technology. A pilot project using the loom is under way in Kanchipuram, a town in the southern state of Tamil Nadu.

Speaking to the online platform Digit at the summit, Muthulakshmi Nellaiappan, who leads AI strategy for Bridgital Loom, noted that a new weaver needs more than 15 years to master the craft. With the technology, a young weaver just entering the trade can become a master far sooner. She also said that because most Indian looms are not electronic, TCS ensured its system works on handlooms too — essential for scaling the technology.

Yet the digital loom raises another question: who owns culture once it becomes pure data? For generations, India's textile knowledge has sat outside the intellectual property system. It lives in community memory, passed from master to apprentice without a certificate of ownership. Kanchipuram, Mekhela-Sador, Sanganeri or Banarasi are the names of places that these fabrics helped shape, not merely places they came from. Now that knowledge is entering generative models — models that owe nothing to these traditions.

For designers and manufacturers, the prospect is alluring: motifs combined at will, colours changed with a single command, thousands of designs generated instantly. The machine brings abundance. The artisan has always lived with scarcity: of money, of bargaining power, and of recognition.

Ananthalakshmi Vadivel, a third-generation weaver who runs Vadivel Silk Looms in Kanchipuram, says her workshop must compete in the market with cheaper Kanchipuram saris made on power looms. With AI, she says, those sellers can produce on-trend designs faster and more cheaply to attract mass consumers, but imitations will never replace the real thing. There will always be a market for authentic Kanchipuram and handwoven fabric, however cheap technology makes the fakes, because the true weave, colour and design of Kanchipuram can only be made by hand — and discerning buyers know it.

Even so, the question is no longer whether AI can reproduce a fabric. It certainly can. The question is: who gets paid when it does?

The legal gap

In December 2025, the federal Department for Promotion of Industry and Internal Trade proposed a statutory "blanket licensing" mechanism under which AI developers could train models on lawfully collected copyrighted material without seeking permission from each rights holder. Once commercialised, such systems would have to pay royalties to rights holders at rates set by the government.

But intellectual property law assumes a traceable work and a traceable author. A motif, by contrast, is the shared property of a community, its technique handed down across generations with no single originator. In such cases, who is entitled to royalties and authorship? The first artisan to create the design? The photographer who recorded it? The researcher who compiled the dataset? The developer who built the tool? Indian law offers no clear answer. The artisan whose name has never appeared on a label may once again be absent from the paperwork.

Meanwhile, the government is planning a dedicated AI research division for handloom textiles at the Centre of Excellence for Handloom Technology at the Indian Institute of Technology (IIT) in New Delhi. The $1.4 million facility, inaugurated in August, aims to modernise handloom textiles by combining traditional textile heritage with contemporary technology. The government also plans to register thousands of weavers, attach sensors to their looms to convert woven patterns into digital data, and issue QR codes for traceability. Maharashtra state's Vingatha system does the same at state level, authenticating handwoven saris such as Paithani and Karvat Kathi.

The idea is simple: if a sari can say where it came from, the person who made it is harder to erase. While such measures cannot stop mass design and production using AI, they can give artisans a clearer presence and turn each piece into an archive.

The name cannot be separated from the community

In galleries, traditional textiles are increasingly shedding the "handicraft" label for the more lucrative category of contemporary art. Artist Gurjeet Singh from Chandigarh recently held a solo textile exhibition in New York, turning fabric scraps, buttons and discarded embroidery into sculptures exploring gender identity and social prejudice. Artist Sayan Chanda, now based in London, works with antique Bengali quilts cut and rewoven to question fixed notions of culture. Artist Sumakshi Singh in Gurugram uses fine embroidery and suspended thread structures to reconstruct fragments of architecture, gardens and personal stories that have disappeared.

In New Delhi, 20 such artisans and artisan-led collectives took part in an exhibition honouring traditional textile crafts. Curator Aparupa Ray Ganguly said the aim was to present these crafts as living practices that continue to change through contemporary artisans, rather than as heritage merely to be archived. One of the participating artists, Kanmai Payun, is the last practitioner of Fulor Saneki — a floral pattern of Assamese weavers. Organisers encouraged her to teach the craft to her 17-year-old son so that it survives after her death.

But as textiles grow more valuable in galleries, their makers do not automatically control their own work. In April, another New Delhi exhibition, titled "Not a Single Weaver", showed artisans absent even at the preview, exposing how traditional textiles are aestheticised while the people who carry that knowledge remain unrecognised.

Chhipa, the artisan and trader in Rajasthan, stresses: "The name cannot be separated from the community. Call your mass-produced products digital prints and leave the traditional names to the authentic artisans."