[Reader Advisory]
This document is currently available in Chinese only. An official English translation has not been scheduled; non-Chinese readers are advised to utilize machine translation for reading purposes.

[License & IP Statement]
The licensing terms of this document equally apply to any non-commercial use of reader-produced translations.
本文件之授權條款同樣適用於讀者自行翻譯後之非商業性使用。

AI Fake Word Art Project: Genesis Record

This document records the genesis and conceptual development of an experimental art project. The project originates from the artist’s acute observation of a specific “error” phenomenon in AI-generated Chinese text images, and from a process of engaging generative AI as a form of “digital prospecting.” Through this interaction, the project gradually establishes its core ideas, formal strategies, and critical objectives. Beyond questioning the nature and influence of AI, the project extends toward a broader critique of how digital information environments simultaneously erode notions of “authenticity” and “quality.”

This record aims to articulate the dynamic evolution of the concept, to frame human–AI collaboration as a mode of “mining,” and to document the artist’s personal reflections shaped by rapid technological shifts. It serves as the foundational document for an artistic inquiry centered on “pollution” and “generation,” situating AI-produced pseudo-characters as cultural and critical material.

《AI 生成假字藝術計畫:概念誕生記錄》

本文件記錄了一項實驗性藝術計畫的概念誕生與發展過程。計畫源於藝術家對當前人工智慧生成中文圖像中特定「錯誤」現象的敏銳觀察,並透過與生成式 AI 進行一場特殊的「數位挖礦」,逐步確立其核心理念、形式策略與最終目標。此計畫不僅探討 AI 的影響,更延伸至對廣義數字信息環境中「真實」與「品質」雙重侵蝕的批判性反思。本記錄旨在呈現概念發展的動態過程、將人機協作視為一種「挖礦」行為的特質,以及藝術家在時代變遷中的個人體驗與反思,並為這項探討「污染」與「生成」的藝術行動提供文獻基礎。

The initial draft of this project was completed in May 2025.
This Zenodo entry presents its first formally published version.

作者背景 | Author Background

作者原職為台灣家具維修匠師與玩具設計師,無工程/研究訓練。2025年5月生成式AI成熟後,7個月內產出2篇DOI論文+20專案,方法源自工藝實作中的AI行為觀察。

Taiwan furniture craftsman & toy designer. No engineering/research training. May-Dec 2025: 2 DOIs + 20 proprietary cases. Methodology from hands-on AI observation in craft workflows.