Google Harnesses AI Technology to Protect and Restore Coral Reefs

Google has introduced an innovative AI tool designed to enhance marine biologists' understanding of coral reef ecosystems and their overall health, contributing to conservation efforts. Named SurfPerch, this tool was developed in collaboration with Google Research and DeepMind. It was trained on thousands of hours of audio recordings from reefs, enabling scientists to "hear reef health from within," monitor reef activity during the night, and assess reefs located in deep or murky waters.

The project began with an engaging initiative that allowed the public to listen to reef sounds online. In just one year, visitors to Google’s Calling in Our Corals website contributed by listening to over 400 hours of reef audio from various global locations, clicking whenever they detected a fish sound. This collaborative effort produced a "bioacoustic" dataset that focuses on coral reef health. By crowd-sourcing this activity, Google built a comprehensive library of new fish sounds, which were essential for fine-tuning the SurfPerch AI tool. As a result, SurfPerch can now be rapidly trained to identify any new reef sounds.

“This advancement enables us to analyze new datasets with unprecedented efficiency, eliminating the need for costly GPU processors while opening new pathways to understanding reef communities and enhancing conservation efforts,” stated a Google blog post that highlighted the project. This blog entry was co-authored by Steve Simpson, a professor of Marine Biology at the University of Bristol in the U.K., and Ben Williams, a marine biologist from University College London, both of whom specialize in coral ecosystems with an emphasis on climate change and restoration.

Additionally, researchers discovered that they could enhance SurfPerch's performance by incorporating recordings of bird songs. Despite the obvious differences between bird and reef sounds, certain common patterns emerged that the model could leverage for learning.

By combining the Calling in Our Corals dataset with SurfPerch during preliminary trials, researchers successfully identified significant differences between protected and unprotected reefs in the Philippines, tracked restoration efforts in Indonesia, and gained insights into fish community dynamics on the Great Barrier Reef.

The project is ongoing, with new audio continuously being added to the Calling in Our Corals website to further train the AI model, according to Google.