Abstract
This paper classifies Al-driven detection methods. compares databases. and determines important areas of research towards protecting lo MT networks. Artificial Intelligence (AI). and more specifically Machine Learning (ML). has come to the fore as a strong solution to improve lo MT security. Al models. by observing traffic patterns, are able to identify anomalies and malicious traffic more accurately and dynamically. This review paper discusses an in-depth overview of the existing threat environment in lo MT networks and investigates diverse Al-based methods for threat detection. It also discusses popular datasets, tools, and frameworks used in the literature. The literature review spans from 2019 to 2025. including both base studies and recent developments. The paper also points out current challenges such as data privacy, memory scarcity, and the demand for real-time detection. Future research needs to concentrate on building explainable, privacy-preserving Al models and standardizing realworld lo MT datasets for real-world deployment.