New AI Spots 63 Tree Species With 95% Accuracy

While artificial intelligence has been watching ground-dwelling animals for years, creatures living high in the trees have been largely invisible to technology—until now.

Scientists have developed TropiCam-AI, a new tool designed specifically to detect and identify animals living in tropical forest canopies across the Americas. The model can now recognize 84 different groups of species, including 63 individual species, with 95% accuracy for most animals it spots.

Andrea Zampetti, a Ph.D. candidate at Sapienza University of Rome, led the project to solve a problem that's been frustrating researchers for years. Existing AI models focus almost exclusively on animals that walk on the ground, leaving the treetop dwellers unstudied despite their crucial role in forest ecosystems.

"We set up TropiCam-AI with the objective of developing a tool that is specifically meant for neotropical camera-trapping surveys targeting the canopy," Zampetti explained. He spent three months in Brazil gathering camera trap footage, working alongside local communities and the Instituto Juruá to capture images of animals in their natural habitat.

These tree-dwelling species matter more than most people realize. Primates, small mammals, and birds living in the canopy consume up to 90% of plant species in tropical rainforests, spreading seeds that keep entire ecosystems alive.

To train the AI, Zampetti collected additional footage from researchers in Peru, Costa Rica, and French Guiana, plus thousands of images from citizen scientists on iNaturalist. The team then manually identified every animal in every photo, teaching the algorithm what to look for.

Why This Inspires

The tool goes beyond simple identification. When an image is too blurry or unclear, TropiCam-AI doesn't force a guess that might be wrong. Instead, it tells users the animal belongs to a particular family or group, maintaining scientific accuracy even when perfect identification isn't possible.

This matters especially now, as deforestation threatens the tree-dependent species the AI was built to protect. By automating the analysis of millions of camera trap images, researchers can track and monitor vulnerable populations much faster than before.

Zampetti and his team continue improving the model with more training data from collaborators across Latin America. "At the end of the day, these kinds of tools are only able to do what you're training them to do," he said.

The technology transforms work that once took months into something that happens in minutes, giving conservationists the speed they need to protect species before it's too late.