When AI meets a 1,200-year-old craft

If you haven’t already, subscribe and join our community in receiving weekly tech insights, updates, and interviews with industry experts straight to your inbox.

Expand your perspective with insights and interviews from the global LEAP community – in your inbox every week. 

This week we’re quoting…

Gareth Hughes (Vice President of IT at Samsonite)

What Hughes said: 

“Our craft heritage gives us a deep understanding of how people travel, what stresses luggage endures, and what ‘trust’ means in a product that literally protects someone’s most valuable belongings. Technology becomes a way to extend that trust.”

From suitcase to kimono

That quote comes from a conversation the DeepFest team had with Hughes about the convergence of heritage craft and tech. And it was the first thing we thought of when we discovered a new heritage/tech experiment happening in Japan. 

Kyoto in winter has a particular quiet to it. Walk through the Nishijin district in the early morning and you can still hear the rhythmic clatter of looms from behind wooden workshop doors – a sound that has echoed through these streets for centuries. 

For more than 1,200 years, the neighbourhood has been the centre of Nishijin-ori, the intricate silk weaving technique used to produce some of Japan’s most elaborate kimono fabrics.

Historically, these textiles were symbols of prestige and ceremony. Patterns drew on nature, mythology and seasonal change – think cranes in flight, maple leaves in autumn wind, waves breaking across stylised clouds. Each design could take thousands of threads and months of work on a loom.

But traditions like this struggle to survive rapid change. 

Over the past few decades, demand for formal kimono has fallen sharply as lifestyles have changed. Workshops that once employed generations of artisans have gradually closed. The average age of master weavers has crept upwards, and skills that used to be passed from parent to child are now at risk of disappearing. 

Which is why, in one of Kyoto’s oldest craft districts, an unlikely collaborator has appeared. Artificial intelligence. 

A neural network has entered the workshop

In a project involving researchers from Sony Computer Science Laboratories and Kyoto’s weaving community, machine learning models have started generating new textile patterns for Nishijin artisans to experiment with. 

The system is trained on archives of traditional motifs and colour palettes, analysing thousands of historical fabric designs to learn how shapes and geometry and colour relationships interact.

Once trained, the AI starts to propose new pattern variations. 

Some suggestions stay close to tradition: rearranged florals, slightly altered geometric symmetries, subtle shifts in colour harmony. Others are stranger – including angular reinterpretations of natural motifs, unusual gradients, or compositions that no traditional design book would normally suggest.

But that’s OK – because the point isn’t to replace the artisan. The idea is to give them something unexpected to respond to. 

Once designers have reviewed a set of AI-generated patterns, they select the ones they feel are most promising or intriguing. 

Those digital concepts then move into the traditional production process, where master weavers interpret them thread by thread on the loom.

Creativity across centuries

For us, the most fascinating element of this collaboration is the way it follows the historical evolution of craft itself. 

Nishijin weaving has always been influenced by outside forces – Chinese silk traditions; European textile techniques introduced in the 19th century; industrial jacquard looms that automated certain weaving processes.

Technology has repeatedly altered the craft without destroying its core.

So in this sense, AI is just the newest visitor to the workshop. But it’s different from earlier innovations in an important way. Traditional design inspiration often comes from observation. A generative system, by contrast, searches through patterns statistically, recombining elements into ideas that are both novel and familiar. 

Artisans describe the experience less as outsourcing creativity and more as receiving suggestions from an apprentice who never stops experimenting.

Some of those suggestions are totally useless – but others are beautiful.

Why preservation sometimes needs invention

There is a deeper logic behind projects like this. Cultural preservation often focuses on protecting traditions exactly as they are, but living crafts don’t survive by remaining frozen in time. They’re more likely to last if they can evolve and adapt to new aesthetics and tastes, and to new economic realities.

For Kyoto’s weaving industry, one of the biggest challenges has been attracting younger audiences who rarely wear traditional kimono. Designers have begun experimenting with contemporary applications: interior textiles, fashion accessories, architectural materials.

AI-assisted pattern generation opens another door. It can produce variations that maintain the spirit of Nishijin design while introducing visual ideas that resonate with modern tastes.

We can think of the tech as a bridge between centuries, then – extending the vocabulary of the past by repurposing it for the present. 

A different vision of AI and creativity

A lot of the public conversation around generative AI focuses on automation – the fear that machines will eventually replace artists, designers or writers. But the Kyoto weaving experiment offers a more collaborative picture of how the technology might actually evolve.

Instead of making creative traditions redundant, AI can act as a catalyst inside them. A source of variation and a generator of possibilities – and most importantly, a partner in exploration. 

In Nishijin workshops, the loom still needs human hands and human patience. The silk still has to be threaded, tensioned and woven with extraordinary care. No algorithm can replicate the tactile knowledge accumulated through decades of practice.

But somewhere upstream, the process has gained a new voice – a model trained on centuries of textile history, that never gets tired of suggesting the next possible pattern. 

AI is a new thread on the loom 

For centuries, Nishijin weaving has been about bringing thousands of threads together into a single pattern. Today, one more thread has joined the loom – an algorithmic one. 

If you liked this, you’ll also enjoy our article about new tech for old cities: read it here. 

If you’re curious about where technology and creativity collide next, those conversations are happening every day across the LEAP ecosystem from 31 August-3 September 2026 – from startup studios to heritage industries experimenting with new tools.

Have an idea for a topic you'd like us to cover? We're eager to hear it. Drop us a message and share your thoughts.

Catch you next week,
The LEAP Team