Arts & Design

AI Meets the Potter's Wheel: Personalized Design in Traditional Ceramic Craft

Traditional ceramic craft faces a tension: growing demand for personalized designs that conventional methods cannot efficiently deliver, alongside concerns that AI-driven design may erode the cultural authenticity that makes handcraft valuable. Recent studies explore both the potential and the risks.

By Sean K.S. Shin
This blog summarizes research trends based on published paper abstracts. Specific numbers or findings may contain inaccuracies. For scholarly rigor, always consult the original papers cited in each post.

Traditional ceramic craftβ€”pottery, porcelain, tile workβ€”carries cultural meaning that extends beyond the functional object. The specific forms, glazes, decorative motifs, and firing techniques of a ceramic tradition encode centuries of accumulated knowledge and aesthetic sensibility. When consumers increasingly demand personalized designs (unique patterns, custom forms, individually decorated pieces), traditional artisans face a production challenge that AI-driven design tools could potentially addressβ€”but not without risk.

The Research Landscape

AI-Driven Personalization

Lei and Wang (2026) investigate how AI can enhance creative capabilities in ceramic art while preserving cultural authenticity. Their approach uses generative models trained on databases of traditional ceramic forms and decorative patterns to produce novel designs that are stylistically consistent with specific traditions while incorporating personal customization (initials, color preferences, dimensional requirements).

The system produces designs that ceramic artisans can then execute by handβ€”using AI as a design aid rather than a manufacturing replacement. Preliminary evaluation by both artisans and consumers found the AI-generated designs acceptable in cultural authenticity and appreciated for personalization, though some artisans expressed concern about the "homogenizing" effect of training dataβ€”the AI tends to produce designs that average the tradition rather than capturing its full expressive range.

Deep Learning for Handicraft

Chen and Lou (2025) present a broader framework for integrating AI into traditional handicraft creation, covering not just ceramics but weaving, woodcarving, and metalwork. Their deep learning system analyzes patterns from existing craft objects and generates new patterns that maintain structural coherence (the generated pattern could physically be woven, carved, or cast) while introducing novel elements.

The key technical challenge is constraint satisfaction: a generated pattern must not only be aesthetically pleasing but physically realizable with the tools and materials of the specific craft tradition. A woven pattern must be loomable; a carved pattern must be chisellable; a ceramic glaze pattern must be achievable with available glazes and firing conditions.

Heritage Education and Commerce

Sari, Ridho, and Arief (2025) address a different dimension: using AI chatbot-based CAD tools to support both education about traditional crafts and digital commerce of craft products. Their prototype, focused on Dinoyo ceramic crafts from Malang, Indonesia, allows users to learn about the craft tradition, explore design variations, and commission personalized pieces through a conversational AI interface.

The motivation is demographic: Generation Z interest in traditional crafts is declining, and interactive digital interfaces may re-engage younger audiences who are comfortable with AI interaction but unfamiliar with traditional craft contexts.

The Risks of Technological Intervention

Yang and Toyong (2026) provide the necessary counterpoint, documenting the unintended consequences of technological intervention in Yazhou potteryβ€”a nationally recognized Chinese intangible cultural heritage. Their analysis reveals that digital tools, while improving production efficiency, also introduce standardization that erodes the variation that characterizes handcrafted objects. The slight irregularities, individual touches, and material responsiveness of handmade pottery are part of its cultural valueβ€”and AI optimization that removes these variations may improve "quality" by engineering metrics while degrading cultural significance.

Critical Analysis: Claims and Evidence

<
ClaimEvidenceVerdict
AI can generate culturally consistent personalized ceramic designsLei & Wang's generative model experimentsβœ… Supported β€” with artisan concerns about homogenization
Deep learning can generate physically realizable craft patternsChen & Lou's constraint-satisfaction framework⚠️ Uncertain β€” demonstrated for simple patterns; complex crafts untested
AI chatbots can re-engage younger audiences with traditional craftsSari et al.'s Gen Z prototype⚠️ Uncertain β€” prototype built; engagement outcomes not measured
Technological intervention can erode handcraft cultural valueYang & Toyong's Yazhou pottery caseβœ… Supported β€” standardization effect documented

What This Means for Your Research

For craft scholars, the Yazhou pottery case provides an empirically grounded warning: technology adoption in craft should be evaluated for cultural impact, not just efficiency. For designers, AI works best as a design assistant (generating options for human artisans to evaluate and modify) rather than a design replacement (fully automating the creative process).

Explore related work through ORAA ResearchBrain.

References (4)

[1] Lei, Z. & Wang, X. (2026). AI-Driven Personalized Design in Traditional Crafts. Journal of Cases on Information Technology.
[2] Chen, Y. & Lou, Y. (2025). Deep Learning-Driven Craft Design: Integrating AI Into Traditional Handicraft Creation. IEEE Access.
[3] Sari, H.V., Ridho, M.H., & Arief, M. (2025). AI Chatbot-Based CAD Prototype to Support Sustainable Heritage Education and Digital Commerce. Proc. ICEEIE 2025, IEEE.
[4] Yang, G. & Toyong, N.M.P. (2026). The Double-Edged Sword: Drawbacks of Modern Technology in the Research and Development of Yazhou Pottery. Malaysian Journal of Social Sciences and Humanities, 11(1).

Explore this topic deeper

Search 290M+ papers, detect research gaps, and find what hasn't been studied yet.

Click to remove unwanted keywords

Search 8 keywords β†’