Illusions in Humans and AI: How Visual Perception Aligns and Diverges
Jianyi Yang, Junyi Ye, Ankan Dash, Guiling Wang · 2025
Read Illusions in Humans and AI: How Visual Perception Aligns and Diverges by Jianyi Yang, Junyi Ye, Ankan Dash, Guiling Wang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-12422.txt.
Description
Jianyi Yang, Junyi Ye, Ankan Dash, and Guiling Wang test AI vision against the illusions that fool human eyes. Illusions reveal that human sight runs on context and assumption, not raw pixels. As machines take on human visual jobs, the question is whether they share our glitches, and what glitches are theirs alone.
Classic color, size, shape, and motion illusions produce mixed results: some effects surface in models, through training or as pattern-matching side effects. Other distortions are purely artificial, like hypersensitivity to single pixels and confident hallucinations. The authors map where the two visual worlds align and split, arguing the split spots weaknesses that safety work must close.
Catalog record
| CREATOR | Jianyi Yang, Junyi Ye, Ankan Dash, Guiling Wang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 31 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-12422