<p>Global Navigation Satellite System (GNSS) is extensively employed in a variety of applications, including smart cities, intelligent transportation systems and location-based services. However, the increasing adoption of GNSS and easy inference has made it a target of many cybersecurity attacks like hijacking, spoofing and sniffing, etc., that may lead to the breach of integrity for trustworthy computing infrastructure. Signal integrity monitoring is one of the main countermeasures, but recent spoofing cases highlight the need for a more intelligent and effective spoofing detection system for ensuring the trustworthy computing infrastructure. Many previous studies focused on individual signal analysis or hardware-based methods and have not adequately utilized the fusion techniques. Using two different types of datasets, this paper proposes an AI-based hybrid approach to detect GNSS attacks using a method that combines features of the k-nearest neighbor technique, dynamic time warping algorithm, and least squares estimation. We utilized spatial, temporal, signal and motion-based features from GNSS signals and device sensor data to detect spoofing attacks. Our evaluation for the proposed approach using <i>F</i>1-score, precision, recall, and accuracy metrics demonstrates an average accuracy of 98.3%, with a maximum recorded accuracy of 99.4%. Furthermore, the performance of the proposed approach is compared with several baseline machine learning models, and it is observed that the proposed approach achieves higher accuracy than the selected baseline models, with comparable time and space complexity.</p>