AI in Commercial Fishing Fleets

Automatic Identification System messages provide vessel identity, location, speed, and direction information. Global Fishing Watch analyzes movement patterns to estimate apparent fishing activity. Its documentation explicitly distinguishes model estimates from precise records of gear deployment. A vessel slowing down and changing direction can produce a false positive, and incomplete records can also hide events. A gap in received AIS messages needs similarly careful interpretation. Receiver coverage and signal interference can produce gaps, so the absence of a received message does not by itself prove deliberate concealment or illegal fishing. Apparent encounters at sea identify movement patterns that may warrant examination; they do not confirm that cargo changed hands. Catch records provide another kind of evidence. NOAA describes electronic monitoring programs that use technologies such as cameras and gear sensors. Computer vision tools are being developed to identify species, estimate dimensions, and count gear from imagery. Their role depends on the fishery and monitoring program, so camera installation should not be described as universally replacing all observers. Habitat tools address a separate question: where species may occur. NOAA's EcoCast example uses observations of target and protected species together with environmental data to estimate relative catch and bycatch probabilities. It supports decisions under uncertainty rather than guaranteeing a bycatch-free trip. Satellite radar adds observations of vessels independent of AIS transmissions. Combining these sources can support investigation, but each observation still has limits. A map should communicate what was measured, what the algorithm inferred, and what additional evidence would be needed to establish compliance.