Chinese AI tool ERNIE Bot Textual Exploration of False Information

Broadcasting and Anchoring School, Communication University of China, Beijing, 100024, China

^* Corresponding author: fuyue@cuc.edu.cn

Abstract

This study provides an in-depth discussion of the application of the Chinese AI tool ERNIE Bot in disinformation detection. First and foremost, the spreading characteristics of false information are analysed, especially the phenomenon of rapid spreading and difficulty in distinguishing authenticity on social media. Then, the performance of ERNIE Bot in grammatical, sentiment, and lexical analyses is investigated in detail, revealing the limitations it faces when dealing with complex disinformation. In order to improve the accuracy of detection, this paper proposes countermeasures to improve the AI detection algorithm, enhance data training and model optimisation, and human-machine collaboration. These countermeasures can not only enhance the detection ability of AI, but also give full play to human judgement in information screening, forming a more effective disinformation prevention mechanism. This study provides valuable theoretical support and practical guidance to enhance the application effect of AI in disinformation detection, aiming to escort social information security.