At Bayside Park in Manly, Brisbane, a small flock of curlew sandpipers—slender, gray-brown shorebirds with delicate curved beaks—huddle on the mudflats, unaware they’re part of a global conservation breakthrough. These critically endangered birds, once vanishing into the noise of habitat loss and climate change, are now at the heart of a quiet revolution: artificial intelligence is learning to find them faster, more accurately, and across landscapes too vast or remote for human eyes alone. Led by The University of Queensland, a landmark study has shown that AI can detect birds in drone imagery with 85% greater speed than manual counting—unlocking a powerful new tool in the fight to save species from extinction.
With over 30 researchers from 11 countries contributing to a dataset of nearly 50,000 birds across more than 100 species, this is the most biologically and environmentally diverse collection of drone-based bird imagery ever assembled. The project, published in Remote Sensing in Ecology and Conservation, wasn’t just about scale—it was about survival. "Birds are going extinct at alarming rates, similar to those seen during the five major mass extinction events in Earth's history," says Professor Richard Fuller from UQ’s School of the Environment. The urgency is real: one in eight bird species is now threatened with extinction. But now, conservationists have a new ally in the sky.
Dr. Joshua Wilson, the study’s lead author, explains that manually counting birds in drone images can take hours, even days. "With AI offering a way to process large numbers of images efficiently, drone surveys can be carried out more often and at larger scales," he says. The AI doesn’t replace field biologists—it liberates them. Instead of wading through wetlands or squinting at thousands of pixels, experts can focus on interpreting data, designing smarter surveys, and pushing for real-world protections. In southern Peru, researcher Cesar Fernandez has already used drone-AI pairing to monitor breeding populations of Chilean flamingos, proving its value in remote, fragile ecosystems.
What makes this effort truly transformative is its openness. The trained AI model is freely available, and the dataset—meticulously labeled to species level—is open access, inviting scientists worldwide to build on it. From the salt pans of Western Australia to the mangroves of Southeast Asia, this tool can scale. It works best for large-bodied birds in open habitats like wetlands, where drones can soar above with minimal obstruction. But the implications ripple far beyond.
As climate change accelerates habitat loss and biodiversity declines, the need for rapid, scalable monitoring has never been greater. This fusion of drone technology and AI isn’t a futuristic fantasy—it’s a working solution, already in flight. And for species like the curlew sandpiper, whose numbers dwindle with each passing season, it may come just in time.