Making AI Agents Evaluate Misleading Charts without Nudging
Fuente:
arXiv
Guardado en:
| Autor principal: | |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866918324343930880 |
|---|---|
| author | Panda, Swaroop |
| author_facet | Panda, Swaroop |
| contents | AI agents are increasingly used as low-cost proxies for early visualization evaluation. In an initial study of deliberately flawed charts, we test whether agents spontaneously penalise chart junk and misleading encodings without being prompted to look for errors. Using established scales (BeauVis and PREVis), the agent evaluated visualizations containing decorative clutter, manipulated axes, and distorted proportional cues. The ratings of aesthetic appeal and perceived readability often remained relatively high even when graphical integrity was compromised. These results suggest that un-nudged AI agent evaluation may underweight integrity-related defects unless such checks are explicitly elicited. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_05662 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Making AI Agents Evaluate Misleading Charts without Nudging Panda, Swaroop Human-Computer Interaction AI agents are increasingly used as low-cost proxies for early visualization evaluation. In an initial study of deliberately flawed charts, we test whether agents spontaneously penalise chart junk and misleading encodings without being prompted to look for errors. Using established scales (BeauVis and PREVis), the agent evaluated visualizations containing decorative clutter, manipulated axes, and distorted proportional cues. The ratings of aesthetic appeal and perceived readability often remained relatively high even when graphical integrity was compromised. These results suggest that un-nudged AI agent evaluation may underweight integrity-related defects unless such checks are explicitly elicited. |
| title | Making AI Agents Evaluate Misleading Charts without Nudging |
| topic | Human-Computer Interaction |
| url | https://arxiv.org/abs/2602.05662 |