<p>The growing dependence on Global Navigation Satellite Systems (GNSS) based aviation systems has exposed vulnerabilities in Automatic Dependent Surveillance-Broadcast (ADS-B), such as susceptibility to interference, jamming, and spoofing attacks. Conventional detection methods often suffer from issues like high false alarm rates, latency issues, and poor adaptability in real-world scenarios. To address these concerns, the current research presents an AI-powered anomaly detection framework that combines Convolutional Neural Network (CNN)-based spatial feature extraction with Gated Recurrent Unit (GRU) based temporal sequence learning, for enhanced real-time detection of GNSS interference in the aviation industry. The proposed CNN-GRU-Based Anomaly Detection and Flight Trajectory Prediction (CNN-GRU ADFTP) takes ADS-B flight telemetry data for training, fusing critical aircraft parameters like latitude, longitude, velocity, altitude, and vertical rate, and demonstrates better accuracy in anomaly detection with fewer false positives compared to traditional machine learning approaches. Geospatial anomaly evaluations also unveil areas of high-risk airspace, highlighting the need for ongoing aviation security monitoring. Although the promising results provided by the suggested framework, challenges like class imbalance and scalability issues persist. Future research will focus on refining model sensitivity, integrating external flight metadata and weather factors, and optimizing edge computing applications for enhancing low-latency processing capabilities in densely populated airspace environments. This study adds to the development of secure and resilient air traffic monitoring systems, thus enhancing aviation safety amidst emerging cyber threats.</p>