Mind the Motions: Benchmarking Theory-of-Mind in Everyday Body Language

TL;DR

This paper introduces Motion2Mind, a new benchmark dataset for evaluating AI's ability to interpret nonverbal cues related to mental states, highlighting current AI limitations in this complex task.

Contribution

It provides a comprehensive video dataset with detailed annotations of nonverbal cues and psychological interpretations, expanding ToM evaluation beyond traditional belief-based tasks.

Findings

AI systems perform poorly on NVC interpretation tasks.

Current models tend to over-interpret in explanation tasks.

Significant performance gap exists between AI and human interpretation.

Abstract

Our ability to interpret others' mental states through nonverbal cues (NVCs) is fundamental to our survival and social cohesion. While existing Theory of Mind (ToM) benchmarks have primarily focused on false-belief tasks and reasoning with asymmetric information, they overlook other mental states beyond belief and the rich tapestry of human nonverbal communication. We present Motion2Mind, a framework for evaluating the ToM capabilities of machines in interpreting NVCs. Leveraging an expert-curated body-language reference as a proxy knowledge base, we build Motion2Mind, a carefully curated video dataset with fine-grained nonverbal cue annotations paired with manually verified psychological interpretations. It encompasses 222 types of nonverbal cues and 397 mind states. Our evaluation reveals that current AI systems struggle significantly with NVC interpretation, exhibiting not only a substantial performance gap in Detection, as well as patterns of over-interpretation in Explanation compared to human annotators.