Detecting ChatGPT Generated Fake Reviews Using Supervised Machine Learning

Customers rely on online reviews to make purchasing decisions. However, the prevalence of fake reviews leads to a challenge in trusting the authenticity of online reviews. The manual detection of these fake reviews is hard and time-consuming. This paper presents an automated fake review detection method using text mining and supervised machine learning. In addition, a new fake reviews dataset called ChatGPT-Fake-It is collected from Google Play and chatGPT. Real reviews about the Facebook mobile application are collected from the Google Play store, while fake reviews are generated using chatGPT. A set of lexical and sentimental features are extracted from the textual reviews to build three classifiers. The experimental results reveal that the decision tree-based method achieved better accuracy, precision, recall, and F1-score than other classifiers.