Garbage in, garbage out: Zero-shot detection of crime using Large Language Models

Abstract: This paper proposes exploiting the common sense knowledge learned by LLMs to perform zero-shot reasoning about crimes given textual descriptions of surveillance videos. We show that when video is (manually) converted to high quality textual descriptions, LLMs are capable of detecting and classifying crimes with state-of-the-art performance using only zero-shot reasoning. However, existing automated video-to-text approaches are unable to generate video descriptions of sufficient quality to support reasoning (garbage video descriptions into the LLM, garbage out).