1 min
Can AI save our oyster reefs? A team of scientists put it to the test
With global oyster populations having declined by more than 85% from historical levels, restoring and monitoring these critical ecosystems is more urgent than ever. But traditional monitoring methods aren’t cutting it. A team of researchers that included the University of Delaware's Art Trembanis have taken a new approach, testing an AI model designed to recognize live oysters from underwater images. The findings? The AI model, called ODYSSEE, was faster than human experts and non-expert annotators, processing in just 40 seconds what took humans up to 4.5 hours. But it wasn’t yet as accurate. In fact, the tool misidentified more live oysters than both groups of human annotators. Still, the team found that ODYSEE has real potential to monitor reefs in real time. Why does this matter? As climate change, pollution and overharvesting continue to pressure coastal environments, more precise and non-invasive monitoring tools like ODYSSEE could become essential to restoration efforts and environmental policy. Trembanis can discuss this new tool and its ability to identify live oysters without disturbing the reef. His expertise in oceanography, engineering and robotics expertise was key to the team's work. The results, published in the journal Frontiers, offer both caution and hope in the race to improve ocean monitoring with emerging technologies. To set up an interview with Trembanis, visit his profile and click on the contact button.
