Multimodal detection improves fake news spotting in generative ai age
Rethinking Multimodal Fake News Detection in the Generative AI Era
Computation and LanguageComputer Vision and Pattern RecognitionMultimedia
Summary
Fake news is becoming trickier to detect because it now often mixes real and AI-generated content. The authors created a new dataset, Weibo26, to better study fake news with mixed content. They designed a system called GAHR that looks at the whole news story and its parts to decide if the news is true and whether AI helped create it. Their approach works well on different tests, detecting both fake news and AI-generated content.
What this means in practice
- •For social media moderation teams: Help identify and verify news posts mixing real and AI-generated content to improve content trustworthiness on platforms.
- •For fact-checking organizations: Enhance fact-checking tools to better distinguish fake news that uses AI-generated elements alongside genuine information.
Authors
Wenbin Shen, Guoxuan Qin, Guangxu Yao, Baodong Wang, Yuanbo Rui, Zhichao Lian
Abstract
Generative content is increasingly entering the production and dissemination of news, transforming fake news from manually fabricated or simply manipulated material into complex forms in which native and generated content jointly participate. Existing multimodal fake news detection research primarily focuses on veracity assessment and rarely characterizes how generativity differences affect the reliability of evidence. In contrast, AIGC detection primarily determines whether content is generated or modified by generative models, but it does not by itself establish whether the underlying news event is true. To bridge the separation between these tasks in data and evaluation, we construct Weibo26, a multimodal fake news detection dataset for generative-content scenarios. On this basis, we propose the Generativity-Aware Hierarchical Reasoning (GAHR) framework, which combines global judgment with local correction so that generativity information participates in news-veracity reasoning. Experiments on multiple existing fake news detection benchmarks and Weibo26 show that GAHR achieves competitive veracity-detection performance while effectively identifying generative content.