TrumorGPT: Query optimization and semantic reasoning over networks for automated fact-checking

Abstract

In the age of social media, the rapid spread of misinformation and rumors has led to the emergence of infodemics, where false information poses a significant threat to society. To combat this issue, we introduce TrumorGPT, a novel generative artificial intelligence solution designed for automated fact-checking. TrumorGPT aims to distinguish "trumors", which are rumors that turn out to be true, providing a crucial tool in differentiating between mere speculation and verified facts. This framework merges machine learning with natural language processing techniques, leveraging a large language model (LLM) with few-shot learning for knowledge graph construction and semantic reasoning. TrumorGPT addresses the "hallucination" issue common in LLMs and the limitations of static training data by incorporating retrieval-augmented generation. This approach involves accessing and utilizing information from regularly updated knowledge graphs that consist of the latest news and information, ensuring that fact-checking of TrumorGPT is based on the most recent data. Accessing updated knowledge graphs greatly enhances the proficiency of TrumorGPT in delivering accurate and reliable information promptly. Evaluating with extensive datasets, TrumorGPT demonstrates superior performance in automated fact-checking. Its ability to effectively conduct automated fact-checking across various platforms marks a critical step forward in the fight against misinformation, enhancing trust and accuracy in the digital information age.