Fake News Detectors are Biased against Texts Generated by Large Language Models

• URL: http://arxiv.org/abs/2309.08674v1
• Date: Fri, 15 Sep 2023 18:04:40 GMT
• Title: Fake News Detectors are Biased against Texts Generated by Large Language Models
• Authors: Jinyan Su, Terry Yue Zhuo, Jonibek Mansurov, Di Wang, Preslav Nakov
• Abstract summary: The spread of fake news has emerged as a critical challenge, undermining trust and posing threats to society. We present a novel paradigm to evaluate fake news detectors in scenarios involving both human-written and LLM-generated misinformation.
• Score: 39.36284616311687
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The spread of fake news has emerged as a critical challenge, undermining trust and posing threats to society. In the era of Large Language Models (LLMs), the capability to generate believable fake content has intensified these concerns. In this study, we present a novel paradigm to evaluate fake news detectors in scenarios involving both human-written and LLM-generated misinformation. Intriguingly, our findings reveal a significant bias in many existing detectors: they are more prone to flagging LLM-generated content as fake news while often misclassifying human-written fake news as genuine. This unexpected bias appears to arise from distinct linguistic patterns inherent to LLM outputs. To address this, we introduce a mitigation strategy that leverages adversarial training with LLM-paraphrased genuine news. The resulting model yielded marked improvements in detection accuracy for both human and LLM-generated news. To further catalyze research in this domain, we release two comprehensive datasets, \texttt{GossipCop++} and \texttt{PolitiFact++}, thus amalgamating human-validated articles with LLM-generated fake and real news.

Related papers

• The Reopening of Pandora's Box: Analyzing the Role of LLMs in the Evolving Battle Against AI-Generated Fake News [30.759761034237563]
  Large language models (LLMs) produce convincing fake news at scale. LLMs are 68% more effective at detecting real news than humans. For fake news detection, the performance of LLMs and humans remains comparable.
  arXiv  Detail & Related papers  (2024-10-25T01:58:29Z)
• From Deception to Detection: The Dual Roles of Large Language Models in Fake News [0.20482269513546458]
  Fake news poses a significant threat to the integrity of information ecosystems and public trust. The advent of Large Language Models (LLMs) holds considerable promise for transforming the battle against fake news. This paper explores the capability of various LLMs in effectively combating fake news.
  arXiv  Detail & Related papers  (2024-09-25T22:57:29Z)
• LLM-GAN: Construct Generative Adversarial Network Through Large Language Models For Explainable Fake News Detection [34.984605500444324]
  Large Language Models (LLMs) are known for their powerful natural language understanding and explanation generation abilities. We propose LLM-GAN, a novel framework that utilizes prompting mechanisms to enable an LLM to become Generator and Detector. Our results demonstrate LLM-GAN's effectiveness in both prediction performance and explanation quality.
  arXiv  Detail & Related papers  (2024-09-03T11:06:45Z)
• ReMoDetect: Reward Models Recognize Aligned LLM's Generations [55.06804460642062]
  Large language models (LLMs) generate human-preferable texts. In this paper, we identify the common characteristics shared by these models. We propose two training schemes to further improve the detection ability of the reward model.
  arXiv  Detail & Related papers  (2024-05-27T17:38:33Z)
• Exploring the Deceptive Power of LLM-Generated Fake News: A Study of Real-World Detection Challenges [21.425647152424585]
  We propose a strong fake news attack method called conditional Variational-autoencoder-Like Prompt (VLPrompt) Unlike current methods, VLPrompt eliminates the need for additional data collection while maintaining contextual coherence. Our experiments, including various detection methods and novel human study metrics, were conducted to assess their performance on our dataset.
  arXiv  Detail & Related papers  (2024-03-27T04:39:18Z)
• Adapting Fake News Detection to the Era of Large Language Models [48.5847914481222]
  We study the interplay between machine-(paraphrased) real news, machine-generated fake news, human-written fake news, and human-written real news. Our experiments reveal an interesting pattern that detectors trained exclusively on human-written articles can indeed perform well at detecting machine-generated fake news, but not vice versa.
  arXiv  Detail & Related papers  (2023-11-02T08:39:45Z)
• Prompt-and-Align: Prompt-Based Social Alignment for Few-Shot Fake News Detection [50.07850264495737]
  "Prompt-and-Align" (P&A) is a novel prompt-based paradigm for few-shot fake news detection. We show that P&A sets new states-of-the-art for few-shot fake news detection performance by significant margins.
  arXiv  Detail & Related papers  (2023-09-28T13:19:43Z)
• MAGE: Machine-generated Text Detection in the Wild [82.70561073277801]
  Large language models (LLMs) have achieved human-level text generation, emphasizing the need for effective AI-generated text detection. We build a comprehensive testbed by gathering texts from diverse human writings and texts generated by different LLMs. Despite challenges, the top-performing detector can identify 86.54% out-of-domain texts generated by a new LLM, indicating the feasibility for application scenarios.
  arXiv  Detail & Related papers  (2023-05-22T17:13:29Z)
• Faking Fake News for Real Fake News Detection: Propaganda-loaded Training Data Generation [105.20743048379387]
  We propose a novel framework for generating training examples informed by the known styles and strategies of human-authored propaganda. Specifically, we perform self-critical sequence training guided by natural language inference to ensure the validity of the generated articles. Our experimental results show that fake news detectors trained on PropaNews are better at detecting human-written disinformation by 3.62 - 7.69% F1 score on two public datasets.
  arXiv  Detail & Related papers  (2022-03-10T14:24:19Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.