Fighting Fire with Fire: Adversarial Prompting to Generate a Misinformation Detection Dataset

TL;DR

This paper introduces an LLM-based method to automatically generate a misinformation detection dataset by creating summaries with controlled factual errors, facilitating scalable training of detection models.

Contribution

It proposes a novel prompting technique to generate labeled misinformation data from trusted news articles, reducing manual annotation efforts.

Findings

Generated datasets improve misinformation detection accuracy

Controlled factual errors enhance model training

LLM-based dataset creation is scalable and effective

Abstract

The recent success in language generation capabilities of large language models (LLMs), such as GPT, Bard, Llama etc., can potentially lead to concerns about their possible misuse in inducing mass agitation and communal hatred via generating fake news and spreading misinformation. Traditional means of developing a misinformation ground-truth dataset does not scale well because of the extensive manual effort required to annotate the data. In this paper, we propose an LLM-based approach of creating silver-standard ground-truth datasets for identifying misinformation. Specifically speaking, given a trusted news article, our proposed approach involves prompting LLMs to automatically generate a summarised version of the original article. The prompts in our proposed approach act as a controlling mechanism to generate specific types of factual incorrectness in the generated summaries, e.g., incorrect quantities, false attributions etc. To investigate the usefulness of this dataset, we conduct a set of experiments where we train a range of supervised models for the task of misinformation detection.