<p>This investigation builds upon scholarly research on AI-integrated anomaly detection in maritime safety operations from 2014 to 2024. The bibliometric analysis analyzes key Countries, Journals, Research themes, Keyword frequency, Co-citation analysis, Impactful authors, and significant collaborations among researchers and relevant institutes for Maritime Anomaly Detection Research. Our findings highlight substantial contributors, including China, France, Italy, and the USA, emphasizing essential collaboration networks and influential journals such as IEEE Transactions on Intelligent Transportation Systems and the International Journal of Approximate Reasoning. We scrutinize the application of various AI and machine learning models, such as LSTM, CNN, and SVM, across different anomaly types. These models are frequently implemented to detect environmental, navigation, cybersecurity, sensor, operational, communication, and behavioural anomalies. For navigation anomalies like unexpected course changes and speed variations, LSTM is utilized due to its capability to handle sequential data. CNN is employed for spatial-temporal pattern recognition in navigation routes, while SVM is applied to classify navigation patterns and detect deviations. Similarly, LSTM effectively identifies security anomalies, including unauthorized boardings and suspicious activities, through sequence modelling, and CNN proves helpful for spatial analysis. This study provides comprehensive insights into the current state of maritime anomaly detection research. It proposes future research areas, emphasizing model accuracy and efficiency enhancements, developing hybrid models, and addressing emerging challenges. Our analysis aims to guide researchers and practitioners by identifying critical areas for advancement and fostering potential collaborative efforts in this field.</p>