TNC EDGE AI Achieves 6% Catch Count Error in Open-Source Fisheries Electronic Monitoring

Title

TNC EDGE AI Achieves 6% Catch Count Error in Open-Source Fisheries Electronic Monitoring

Content

The Nature Conservancy (TNC) released its full AI-powered electronic monitoring solution as open source to advance ocean conservation, combat illegal fishing, and usher in a new era of data-driven fisheries management, achieving a 6% catch count error rate compared to expert human reviewers. "For electronic monitoring to truly work at a global scale, it has to be accessible, affordable and fast," said Ben Gilmer, director of large-scale fisheries at The Nature Conservancy. "That's why we're making this technology freely available — to allow governments, scientists and industry to build from the same foundation and accelerate a future where verified on-the-water activities and accurate fisheries data are not just an aspiration, but a global standard critical to science-based management, market access and combating illegal, unreported and unregulated fishing." The context was that industrial fishing takes place across more than half of the ocean's surface — an area three times larger than all the world's farmland — and supplies seafood to billions of people, yet nearly 90% of global marine fish stocks are fully exploited, overexploited or depleted. For decades, fisheries managers lacked the timely and verifiable data needed to manage increasingly complex fishing operations. Catch reporting has largely been unverifiable, creating conditions for licensed vessels to engage in illegal, unreported and unregulated (IUU) fishing. Technically, the system uses edge computing to analyze EM video directly onboard longline fishing vessels equipped with Starlink terminals, with computer vision models detecting, tracking, and classifying catch as it is brought onboard. The system compares AI predictions to captains' electronic logbooks each day and flags potential non-compliance before vessels return to port. The full solution stack was built by Tryolabs, a Uruguayan AI company. The prototype was selected as one of 15 global awardees in the Bezos Earth Fund's AI for Climate and Nature Grand Challenge, with the $2 million award supporting scaling to the Western and Central Pacific Ocean, including The Republic of Palau. At scale, the system runs onboard vessels continuously rather than relying on delayed shore-based review. By making the technology freely available, TNC aims to enable governments, scientists, and industry to build from the same foundation and accelerate global adoption.