AI-Driven Online Fake News Detection System in the Healthcare Industry

The digital age has facilitated the rapid dissemination of information, including health-related content. However, this has also led to the widespread spread of fake news, damage public trust in medical institutions and threatening public health. Traditional methods of detecting health fake news are often insufficient in addressing the scale and complexity of online content. AI-based systems capable of automatically recognizing and classifying health misinformation in real-time are urgently needed. Additionally, Z-Score Normalization evaluates unknown data, where a missed prediction signifies a lost performance value. Furthermore, the Grey Wolf Optimization (GWO) Algorithm mimics the hunting behavior and social hierarchy of grey wolves. The algorithm divides the solution population into three categories: real-time applications, testing on historical data, and training using current data, along with verification for each dataset. The analysis of calculations in the normalized performance data lack’s reliability regarding performance. This paper proposes an AI-enabled misinformation detection framework based on Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) networks to recognize false healthcare content across various online platforms. The system processes articles, social media threads, and health blogs and applies Natural Language Processing (NLP) tools for data preprocessing, feature selection, and classification. The long-term dependencies are better handled by LSTM networks, which have computationally efficient operations that enable the model to learn subtle patterns for identifying online fake news. Performance metrics (precision, recall, F1-score, and ROC-AUC) demonstrate the model's efficiency, with performance metrics exceeding 98% for all indicators. The AI-powered model is used to address the critical problem, thereby creating an effective means for detecting misleading health rumours in real-time, allowing users and media to detect truth from falsehood.