AI RESEARCH
Consistent Yet Wrong: Evidence Insensitivity in Spatial Vision-Language Models
arXiv CS.CV
•
ArXi:2606.02742v1 Announce Type: new Spatial reasoning is fundamental to robotics, autonomy, and embodied AI, yet modern vision-language models (VLMs) remain unreliable on metric distance queries. A common assumption is that consistent predictions across viewpoints reflect geometric grounding. We test this assumption and find the opposite: leading VLMs often produce view-invariant and consistent answers even when those answers are incorrect, indicating weak coupling between predictions and viewpoint-specific visual evidence.