Passive acoustic monitoring (PAM) offers a noninvasive approach for underwater monitoring at large spatial and temporal scales, including in hardly accessible areas, yet it remains underutilized in fish ecology. Many fish produce various sounds associated with ecologically relevant behaviors providing valuable indicators. PAM for fish monitoring is constrained by two major challenges: the limited knowledge and availability of reference sounds in libraries for species identification and the absence of efficient, standardized data analysis methods, resulting in a dependence on labor-intensive manual analyses. Recent advances in machine learning, particularly deep learning, offer promising solutions for automated detection and classification of fish sounds, improving data analysis efficiency, scalability, and cost-effectiveness. While initial applications demonstrate clear benefits, widespread implementation remains limited. This chapter outlines a road map for large-scale adoption of PAM in fish ecology, emphasizing open datasets, benchmark collections, accessible analytical tools, and expanded reference sound libraries. Addressing PAM’s interconnected main challenges will create a positive feedback loop, where improved methods accelerate data collection and, in turn, enhance methodological development. Ultimately, PAM can become a scalable, flexible, and powerful tool for fish monitoring, conservation, and management in a changing ocean.