UWSNs support oceanographic data acquisition, environmental sensing, offshore infrastructure inspection, and underwater surveillance, but face severe bandwidth, propagation delay, topology, and energy constraints that limit traditional networking. This paper reviews AI-based techniques—machine learning, deep learning, and reinforcement learning—in UWSNs, examining their impact on routing, energy conservation, reliability, localization, and security. A structured comparison identifies performance trends across the literature, and the paper closes with key challenges and future research directions for scalable, intelligent underwater sensing.

Keywords: Underwater Wireless Sensor Networks, Artificial Intelligence, Machine Learning, Deep Learning, Reinforcement Learning, Intelligent Routing, Energy Optimization, Internet of Underwater Things