Researchers develop ensemble learning methods to distinguish AI-generated fake news from human-written misinformation using linguistic and emotional analysis

Study analyzes linguistic, structural, and emotional differences between AI-generated and human-written fake news to understand how deceptive content differs

Multiple classification models tested including logistic regression, random forest, support vector machines, extreme gradient boosting, and neural networks

Document-level features extracted from sentence structure, lexical diversity, punctuation patterns, readability indices, and emotion detection (fear, anger, joy, sadness, trust, anticipation)

Research addresses growing challenge posed by large language models creating new class of AI-generated misinformation alongside traditional human-written fake news