Real-time Fake News from Adversarial Feedback

2024-10-18

arXiv.org

AI Summary

Existing fake news detection methods overly rely on static pattern matching and lack rigorous evaluation of models’ dynamic factual reasoning capabilities. Method: We propose an iterative adversarial generation framework guided by Retrieval-Augmented Generation (RAG) feedback: GPT-4o rewrites real news across multiple rounds, while retrieval-augmented fact verification and ROC-AUC–driven adversarial optimization jointly generate highly deceptive fake news that evades mainstream LLM-based detectors. Crucially, RAG is employed *bidirectionally*—both for detection and for adversarial generation—establishing a closed-loop system centered on real-time factual reasoning. Contribution/Results: This is the first work to unify RAG in both detection and generation for systematic robustness evaluation. Experiments show our method reduces the ROC-AUC of strong RAG-based detectors by 17.5 percentage points, exposing critical vulnerabilities of purely generative detectors under unseen events and adversarial attacks. Our framework establishes a new benchmark and analytical paradigm for robust fake news identification.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Adversarial Learning & Robustness
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

We show that existing evaluations for fake news detection based on conventional sources, such as claims on fact-checking websites, result in high accuracies over time for LLM-based detectors -- even after their knowledge cutoffs. This suggests that recent popular fake news from such sources can be easily detected due to pre-training and retrieval corpus contamination or increasingly salient shallow patterns. Instead, we argue that a proper fake news detection dataset should test a model's ability to reason factually about the current world by retrieving and reading related evidence. To this end, we develop a novel pipeline that leverages natural language feedback from a RAG-based detector to iteratively modify real-time news into deceptive fake news that challenges LLMs. Our iterative rewrite decreases the binary classification ROC-AUC by an absolute 17.5 percent for a strong RAG-based GPT-4o detector. Our experiments reveal the important role of RAG in both detecting and generating fake news, as retrieval-free LLM detectors are vulnerable to unseen events and adversarial attacks, while feedback from RAG detection helps discover more deceitful patterns in fake news.

Problem

• Large Language Models
• Adversarial Fake News
• RAG Technology

Innovation

• RAG Technology
• Adversarial Feedback
• Fake News Detection