Detecting Artificial Intelligence-Generated Textual and Image Misinformation Using Machine Learning

Artificial Intelligence (AI)-generated media misinformation poses a significant threat to public trust, societal stability, and legitimacy online. Rapid advancements in AI-generated media in the forms of text and image misinformation, have made it increasingly accessible for the public to create media that can deceive and manipulate public opinion on a large scale. This study provides a comprehensive review of the current methodologies for detecting AI-generated textual and image misinformation, identifying key challenges in the field, and highlighting opportunities for further exploration to enhance detection capabilities. Key to this study are natural language processing and convolutional neural network techniques, which provides a framework for developing tools for the identification and detection of AI-generated textual and image misinformation. Notably, RoBERTa-based models have demonstrated an accuracy of up to 99.2% in detecting AI-generated tweets, underscoring the potential of advanced machine learning techniques in addressing this growing issue. By analyzing real-world examples, this review dives into the application of NLP and CNN for AI-generated media misinformation and provides a comprehensive overview of the progress of machine learning in detecting AI -generated media misinformation.