发表机构
KU Leuven; TU Darmstadt; National Research Center for Applied Cybersecurity ATHENE; Mohamed bin Zayed University of Artificial Intelligence（鲁汶大学; 达姆施塔特工业大学; 国家应用网络安全研究中心ATHENE; 穆罕默德·本·扎耶德人工智能大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对AI生成图像与声明不一致的虚假信息，提出MIC框架，结合SFT和GRPO优化，并构建MIC-Bench基准，显著提升核查性能。
AI 中文摘要
与AI生成图像配对的声明是一种快速增长的虚假信息形式。现有的自动化事实核查（AFC）方法主要将此视为溯源问题，通过检测低级合成伪影来判断图像是否为AI生成。然而，这些方法并未验证人类事实核查者经常检查的内容：图像内容是否与伴随声明所暗示的上下文一致。为解决这一差距，我们引入了MIC（多模态不一致性检查），一个AFC框架，通过检测AI生成的多模态虚假信息并利用世界知识解释不一致性来辅助人类事实核查者。MIC首先使用监督微调（SFT）进行任务适配，然后应用组相对策略优化（GRPO）直接优化组件级可验证奖励，用于裁决预测、不一致类型分类、视觉证据描述和世界知识解释。我们进一步引入了MIC-Bench，一个包含8,812个图像-声明实例的基准，这些实例源自4,406个声明，每个声明配有一个真实图像和一个引入受控上下文不一致性的AI生成对应图像。与仅使用SFT相比，GRPO在分布内和分布外设置中分别将Macro-F1提高了4.67和4.11个百分点，同时提高了视觉证据描述和世界知识解释与参考注释的语义相似性。我们的代码和数据可在以下https URL获取。
英文摘要
Claims paired with AI-generated images are a rapidly growing form of misinformation. Existing automated fact-checking (AFC) methods mainly treat this as a provenance problem, detecting low-level synthesis artifacts to decide whether an image is AI-generated. However, such methods do not verify what human fact-checkers often check: whether an image's content is consistent with the context implied by its accompanying claim. To address this gap, we introduce MIC (Multimodal Inconsistency Checking), an AFC framework that assists human fact-checkers by detecting AI-generated multimodal misinformation and explaining inconsistencies using world knowledge. MIC first uses supervised fine-tuning (SFT) for task adaptation and then applies Group Relative Policy Optimization (GRPO) to directly optimize component-level verifiable rewards for verdict prediction, inconsistency type classification, visual evidence description, and world-knowledge explanation. We further introduce MIC-Bench, a benchmark comprising 8,812 image-claim instances derived from 4,406 claims, where each claim is paired with an authentic image and an AI-generated counterpart that introduces a controlled contextual inconsistency. Compared with SFT alone, GRPO further improves Macro-F1 by 4.67 and 4.11 points in the in-distribution and out-of-distribution settings, respectively, while also improving the semantic similarity of visual evidence descriptions and world-knowledge explanations to reference annotations. Our code and data are available at https://github.com/UKPLab/arxiv2026-mic.
CommentsPreprint under review
