Researchers build first AI tool capable of identifying individual birds

In animal behaviour studies, individually identifying animals is one of the most expensive and time-consuming factors, which limits the scope of behaviours and the size of the populations that researchers can study. The ability to distinguish individual animals from each other is also important for the long-term monitoring of populations and the protection of species from pressures such as climate change.

In this study, the authors demonstrate that artificial intelligence can be used to identify individual birds. They show that computers, when trained, can consistently recognise dozens of individual birds, even though we cannot, as humans, tell these individuals apart.

This finding is a great step forward for the development of methods for automatic, non-invasive identification of animals which are completely unmarked and unmanipulated by researchers.

To see some examples of the news please click on the links below:

"Not even scientists can tell these birds apart. But now, computers can” | Science Magazine| July 28, 2020
"AI model developed to identify individual birds without tagging” | The Guardian | July 27, 2020
"Birdwatching AI can recognise individual birds from behind” | NewScientist | July 27, 2020"
"AI software identifies individual wild birds without need for ringing or tagging” | Independent | July 27, 2020
"AI software which recognises individual birds could boost conservation efforts, researchers say” | The Telegraph | July 27, 2020
"Inteligência artificial "treinada" para identificar aves individualmente” | SIC Notícias | July 27, 2020
"Investigadores usam inteligência artificial para identificar aves individualmente" | Observador | July 27, 2020

To read the national press release for this study, please click here.

To access the original article, please click here.