Lies We Can See: Joint Verbal and Non-Verbal Deception by VLM Agents in Embodied Social Interactions
MineAmongUs is introduced, a 3D multimodal Among Us sandbox where imposter agents must deceive crewmates through joint verbal and non-verbal action, and ARIA is proposed, a configurable VLM-agent harness that exposes five cognitive-component ablation axes and opens a new path for embodied VLM-agent alignment research.
Jaewoo Ahn, Junseo Kim, Hyunseo Kim et al.
· 0 citations