Arita ware, porcelain made in Arita Town and surrounding areas of Saga Prefecture, is produced through a division of labor, with different people responsible for each stage. To address the challenge of passing the knowledge and judgment involved in each stage on to the next generation, AI-assisted recordkeeping, such as transcribing audio, could be one option. This article looks at how AI can serve not as a substitute for craftspeople, but as an aid to documenting their work.
Why Consider Traditional Crafts and AI Together?
Many traditional crafts have been handed down from person to person over long periods of time. If that transmission is interrupted, the shape and patterns of a vessel may remain, but the reasoning behind how and why it was made that way can be lost. Passing on these skills is an ongoing social challenge for craft-producing regions.
AI is often portrayed as a threat that will take work away from craftspeople. But replacing everything with machines is fundamentally different from using technology to help preserve people’s words and experience. The latter should be the starting point when considering the relationship between traditional crafts and AI.
Arita ware is featured here because its division-of-labor structure offers a clear example for exploring how to document craft skills. In a production region where each stage has its own specialists, the knowledge associated with an entire stage could disappear if one person leaves. That is why the question of what to preserve, and how, becomes so concrete.
Arita Ware Is Made Through the Work of Many Hands
Arita ware is porcelain made in Arita Town and surrounding areas of Saga Prefecture. Production generally involves a division of labor, with specialists bringing their knowledge and skills to each stage, including forming, decorating and firing. A finished vessel reflects the judgment of several people.
Take Care When Dating Its Origins
The origins of Arita porcelain are sometimes dated to “1616.” However, the Arita Town Historical and Folklore Museum explains that this date comes from a document recording the relocation of Sanbei Kanegae, and cannot be directly linked to the start of porcelain production. If giving a date, it is more appropriate to refer broadly to the “1610s.”
How have skills been passed down in a region said to have a history of more than 400 years? In the following sections, we consider how far the modern tool of AI can contribute.
How to Document the Skills Behind a Division of Labor
When production is divided into stages, passing on the work must also be considered one stage at a time. What a craftsperson responsible for forming looks for is likely different from what a decorator pays attention to. If everything is lumped together under the label “craft skills,” important details can be lost.
Preserve the Reasons Behind the Decisions
What does a craftsperson look for? When do they pause, and why do they make a particular choice? Documenting the reasoning behind these decisions is the starting point for creating a record. Videos can show the steps involved, but the basis for a decision remains invisible unless the craftsperson explains it.
What follows is a proposal for preserving knowledge across the region, not a report on a system introduced at a specific kiln.
Audio Transcription Can Be a Starting Point for Recordkeeping
A practical starting point is to transcribe audio from interviews. OpenAI’s API documentation still lists Whisper as an option for audio transcription. Transcribing interviews with veteran craftspeople or conversations during breaks in the work can make them searchable and easy to revisit later.
For example, imagine a young craftsperson at a workshop asking a senior colleague, “Why did today’s firing turn out this way?” and recording and transcribing their exchange. After several years, the workshop’s accumulated rules of thumb could be preserved in written form. Information once reliant on memory or paper notebooks could potentially become a shareable asset.
Transcription alone, however, cannot reproduce a craftsperson’s feel for the work. Subtle differences in heat control and the feel of the material in one’s hands are partly lost the moment they are put into words. Text is not the craft itself; it is a map that can help people find their way to it.
Deciding What to Entrust to AI
Helping with recordkeeping and reproducing the craft itself are separate challenges. The former is closer to what AI does well; the latter belongs to the realm of experience built up by people. If this distinction is blurred, people in the field will not be convinced.
Ideas such as using generative AI to propose designs for new vessels or AI agents to support sales overseas are conceivable. But these are examples at the conceptual stage, not practices that can be described as established in Arita ware workshops. If such ideas are tested, the sensible first step is for the craftspeople themselves to decide what they want to preserve and what they want to keep in human hands.
The value of using AI lies not in adopting it for its own sake, but in deciding who can use the knowledge of the people in the field, and how it should be preserved for them. Arita ware craftsmanship is the main focus; AI plays a supporting role. In a region where skills have been refined through a division of labor, it makes sense to see AI as another role within that production process.
Comments