Researchers have created an artificial intelligence system that could render ankle monitors for birds obsolete. Trained on thousands of images of birds, the AI searches for unique characteristics in the birds’ feathers that allow it to differentiate one from the other. It can successfully pick out an individual bird from multiple different photos.
When used in a trial in the wild, the system was found to have an accuracy rate of over 90 percent. Because it is completely non-invasive, this system could eliminate the need for attaching physical trackers to wild birds, which would be highly beneficial to both wildlife biologists and the animals themselves.