Detecting Cross-Modal Inconsistency to Defend Against Neural Fake News

TL;DR

This paper introduces a new approach to detect neural fake news by identifying visual-semantic inconsistencies in machine-generated articles, addressing a more complex and realistic scenario than previous methods.

Contribution

It presents a novel dataset of machine-generated news articles with images and captions and proposes an effective detection method based on visual-semantic inconsistency analysis.

Findings

Human studies reveal vulnerabilities in neural fake news detection.

The proposed inconsistency detection method outperforms baseline approaches.

The NeuralNews dataset enables future research in this area.

Abstract

Large-scale dissemination of disinformation online intended to mislead or deceive the general population is a major societal problem. Rapid progression in image, video, and natural language generative models has only exacerbated this situation and intensified our need for an effective defense mechanism. While existing approaches have been proposed to defend against neural fake news, they are generally constrained to the very limited setting where articles only have text and metadata such as the title and authors. In this paper, we introduce the more realistic and challenging task of defending against machine-generated news that also includes images and captions. To identify the possible weaknesses that adversaries can exploit, we create a NeuralNews dataset composed of 4 different types of generated articles as well as conduct a series of human user study experiments based on this dataset. In addition to the valuable insights gleaned from our user study experiments, we provide a relatively effective approach based on detecting visual-semantic inconsistencies, which will serve as an effective first line of defense and a useful reference for future work in defending against machine-generated disinformation.