Illegal fishing poses an existential threat to global fisheries and is responsible for causing significant damage and depletion to fish stocks, food security and national economies. We describe a method through our agent-based high-level information fusion system to automate the detection of illegal fishing, thereby providing analysts with real-time indicators, warnings, and a question and answering capability. By automated reasoning over fused human-and sensor-derived sources we demonstrate the capacity to automatically infer candidate cases, which may enable authorities to take action in a timely fashion.