404 Snow
We take digital “bugs” as a starting point, using regional landscape differences and AI-generated snowman images to materialize algorithmic biases. These imperfect outputs offer viewers access to a “second-hand reality” shaped by technological limitations, while preserving space for speculation.
In the warm Bay Area of China, snow is uncommon. We collected descriptions of snowmen from residents who have never seen snow and used them to train the AI, generating 205 images. Rather than accurate representations, they are composites of algorithmic inference and human imagination. As an emerging cognitive apparatus, AI still operates with limited understanding, yet its errors and distortions open unexpected avenues of imagination—producing snowmen whose vitality often exceeds that of precise simulations.

We meticulously organized and refined our results and processes, ultimately unifying them in our publications, including a series of digital snowman posters and questionnaires, as well as a thread-bound book of 205 snowman image feedbacks. We also 3D-printed our snowmen and displayed them against the backdrop of Bay Area sand.



Time:
2025.06.

Category:
Experimental



Designer:
Yao Xiaodan, Wangzhen Jixing, Wang Changqiao
姚晓丹, 王臻吉欣, 王昶乔










































More Works