PrismAgent: Illuminating Harm in Memes via a Zero-Shot Interpretable Multi-Agent Framework

2026-05-01

AI Summary

Existing approaches to harmful meme detection rely heavily on large-scale annotated data, which is costly to acquire and exhibits limited generalization. This work proposes the first interpretable multi-agent framework designed for zero-shot settings, simulating a criminal investigation pipeline through four collaborative stages: analysis, investigation, prosecution, and judgment. By integrating multi-agent collaboration, intent paraphrasing, contextual evidence retrieval, and multi-perspective evaluation, the framework enables explicit, end-to-end reasoning for harmful meme identification. Evaluated on three public benchmarks, the method substantially outperforms current zero-shot alternatives while providing clear, traceable intermediate reasoning steps that enhance interpretability and trustworthiness.

Abstract

The rapid spread of memes makes harmful content detection increasingly crucial, as effective identification can curb the circulation of misinformation. However, existing methods rely heavily on high-volume annotated data, which leads to substantial training costs and limited generalization. To address these challenges, we propose PrismAgent, a zero-shot, multi-agent, interpretable framework. PrismAgent conceptualizes this task as a criminal case investigation, employing four specialized agents responsible for the analysis, investigation, prosecution, and judgment stages within a structured collaborative workflow. In the first stage, the analyst agent paraphrases each meme under benevolent and malicious assumptions to probe its underlying intent. The investigator agent then retrieves supporting evidence from an unannotated dataset and constructs contextual interpretations for the meme and its variants. Next, the prosecutor agent performs three independent preliminary judgments by pairing the original meme with each of the three interpretations. Finally, the judge agent deliberates across all evidence to render a final verdict. Moreover, PrismAgent's explicit multi-stage reasoning chain makes the model inherently interpretable, as every intermediate step is explicitly explained rather than only producing a final detection result. Extensive experiments on three public datasets show that PrismAgent significantly outperforms existing zero-shot detection methods.

Problem

• harmful meme detection
• zero-shot learning
• interpretable AI
• multi-agent framework
• misinformation

Innovation

• zero-shot
• multi-agent
• interpretable
• meme harm detection
• collaborative reasoning