AI assisted craft design should begin with artisan constraints

For founders working with craft communities, design assistance should support a relationship with artisans rather than treat their knowledge as a style library available for unrestricted imitation.

Establish the design boundary with the makers

Discuss materials, tools, production time and elements the community or artisan does not want reproduced. Clarify the intended commercial use and obtain appropriate permissions and professional advice where needed.

UNESCO’s intangible-cultural-heritage convention recognises the role of communities in identifying their heritage. That principle supports involving knowledge holders rather than assigning authenticity through an AI output.

An illustrative enterprise-gifting concept might be visually attractive but impossible to weave at the proposed size. The artisan’s assessment should shape the next iteration.

Separate concept from production evidence

Label generated images as concepts. Do not present them as completed artisan products or imply that a community endorses the design without confirmation.

Prototype with the relevant maker and record the changes required. Material use, finish and repeatability may matter more to the buyer than the detail visible in a rendered image.

Keep attribution and commercial arrangements explicit rather than assuming design speed resolves the relationship between contributors.

Evaluate the collaboration

Measure usable concepts, prototype effort and the artisan’s assessment of fit. A large number of generated variants is not evidence of better product development.

Keep a prototype record that identifies the generated concept, artisan revisions and production test. This can make contribution and feasibility visible to a buyer without implying that the initial image was a finished craft object. The record should follow the permissions and commercial arrangements agreed with the people involved.

GAT’s SEATCE coverage introduces the importance of regional craft knowledge. AI can contribute to a modern product workflow when that knowledge remains connected to the people who practise it and have a voice in how it is used.