Drones and AI Speed Up Bird Conservation by 85 Times

More than half the world's bird species are disappearing, but a University of Queensland researcher just made protecting them vastly easier.

Joshua Wilson traded his engineering career designing drones for something more meaningful: using those same drones to save birds from extinction. Over four years, he developed AI software that counts birds in aerial images 85 times faster than manual counting.

The breakthrough came at a critical moment. More than 200 Australian bird species now face extinction, nearly double the number from the mid-1990s. Even common backyard birds like willy wagtails are declining by 2 to 3 percent annually.

Wilson tested his approach at a shorebird sanctuary near Brisbane, where critically endangered far eastern curlews rest after flying 13,000 kilometers from Siberia. He learned exactly how high and far to fly drones without disturbing the exhausted travelers below.

The AI transforms work that once consumed days into overnight processing. Conservationists can now deploy drones in remote locations, let computers analyze the footage, and spend their expertise on actual conservation instead of tedious counting.

The Ripple Effect

Wilson knew local data wouldn't cut it for a global crisis. He emailed researchers worldwide, and 33 scientists from 11 countries responded with images. The dataset now includes nearly 50,000 birds from over 100 species: penguins in Antarctica, cormorants nesting in African treetops, white storks perched on Polish rooftops.

The diversity makes the AI adaptable anywhere waterbirds gather without tree cover blocking the view. Mediterranean gulls in Italy, chinstrap penguins in frozen southern waters—the system identifies and counts them all.

BirdLife Australia president Hugh Possingham explained why accurate counts matter so much. Australia has lost 25 bird species since European settlement, and testing conservation strategies requires precise population data. With a continent the size of the United States but 20 times fewer scientists, Australia desperately needs tools that work at scale.

Professor Richard Fuller emphasized that fast, scalable monitoring makes the difference between effective conservation and guesswork. Remote locations and difficult terrain have long prevented accurate population tracking.

Both the massive dataset and the AI model are now freely available online. Wilson hopes researchers will keep contributing images, expanding the tool's capabilities with each new species and habitat added.

The innovation transforms an impossible task into manageable science, giving endangered birds a fighting chance as their numbers continue falling worldwide.