AI can track a single bird over time — based solely on its song.
This approach could provide scientists with a faster, gentler way to assess the health of bird populations.
Forests across North America have seen a drop in the number of songbirds that herald the arrival of spring. To find out why, biologists scour forests to capture birds and mark them with bands. When they return a year later to look for the marked birds, they can get an idea of how many survive from one year to the next. But field work takes countless hours, and trapping and banding can stress birds.
A paper published this week in Nature Methods presents a faster, gentler approach. It uses sound recorders and artificial intelligence (AI) to identify individual birds based solely on their song and track them over time. "This is almost the Holy Grail," says Scott Sillett, a biologist at the Smithsonian Institution's Migratory Bird Center who was not involved in the research.