Global Fishing Watch is a nonprofit founded by the ocean advocacy group Oceana, the satellite-analysis nonprofit SkyTruth, and Google to make commercial fishing activity visible and accountable. Its core product is an open, interactive map of where the world’s fishing vessels operate, built by applying machine learning to billions of position signals and satellite images.
The primary data source is the Automatic Identification System (AIS), the radio beacons that ships broadcast with their position, identity, course, and speed. More than 400,000 devices feed the system. Global Fishing Watch’s models classify vessels into roughly 40 categories - trawlers, longliners, and so on - and, crucially, infer when a vessel is actually fishing rather than merely transiting, based on its movement patterns. The same models flag suspicious behavior such as false identities or GPS spoofing. Because only a fraction of fishing vessels carry AIS, the organization layers in satellite imagery: nighttime VIIRS lights, optical imagery, and synthetic aperture radar from the European Space Agency’s Sentinel-1, processing petabytes of radar data to detect vessels that broadcast no signal at all.
That “dark fleet” work culminated in a 2024 study, published in Nature, that combined radar, optical imagery, and AI to reveal that a large share of industrial fishing and offshore infrastructure was missing from public tracking entirely. All of Global Fishing Watch’s datasets are published free through APIs and a public portal, used by governments, researchers, and journalists to monitor protected areas and investigate illegal fishing.
Why business readers should care: Global Fishing Watch shows how AI applied to cheap, ambient signals - radio beacons and free satellite radar - can create regulatory transparency in a domain that was previously invisible, shifting power toward enforcement and away from operators who relied on the ocean being unwatched.