<p>Technological advances such as electronic charts and course prediction systems provide invaluable support to navigation officers and maritime pilots in navigating confined waters. However, recent maritime accidents have been attributed to operators lacking a clear understanding of how the automation works and how to use it, leading to both misuse and disuse. Two concerns emerge: inadequate training and poor automation design, making it difficult and complex to use. To investigate challenges related to understanding, use, and trust in automation, we surveyed Swedish maritime pilots to investigate their experiences with the course predictor automation tool on their portable pilot units. This technology predicts ship trajectories and is commonly used in modern bridge systems. This paper contributes empirical evidence on how maritime pilots trust the predictor currently used, providing insight into their perceptions and experiences of training, level of understanding, and patterns of usage. The results of 69 respondents revealed limited formal training in the predictor, with knowledge acquired primarily from self-learning and practical experience. Although pilots value the predictor and use it frequently, they struggle with sensor error detection and understanding how it works. The trust in the predictor was inversely correlated with age and experience, with lower age and experience associated with higher trust, more frequent use, greater perceived importance, better understanding, and fewer unexplained behaviours encountered. Based on these findings, recommendations are proposed to improve predictor training and improve its transparency through design.</p>