A new artificial intelligence application called DinoTracker can analyze photos of fossilized dinosaur tracks to identify the species that made them, matching expert accuracy in many cases. The tool has revealed potential bird-like footprints dating back over 200 million years, challenging ideas about avian evolution. It also provides fresh insights into mysterious tracks from Scotland's Isle of Skye.
Paleontologists have long struggled to interpret dinosaur footprints, which offer clues about ancient behaviors but often degrade over time. A team from Germany's Helmholtz-Zentrum Berlin and Scotland's University of Edinburgh has developed DinoTracker, an AI-powered mobile app that simplifies this process.
Users upload images of tracks, and the app examines features like toe spread, heel placement, ground contact area, and weight distribution. Trained on about 2,000 actual fossils and millions of simulated variations to account for preservation changes, the system identifies eight distinguishing traits. It achieves roughly 90% agreement with human specialists, even for tricky specimens.
The research, detailed in a 2026 Proceedings of the National Academy of Sciences paper, highlights unexpected discoveries. Some tracks exceeding 200 million years in age show resemblances to bird feet, suggesting birds may have evolved far earlier than thought—or that certain dinosaurs independently developed similar structures.
In Scotland, the app analyzed footprints from the Isle of Skye, preserved in lagoon mud around 170 million years ago. These appear to belong to early relatives of duck-billed dinosaurs, marking some of the oldest known examples globally.
Dr. Gregor Hartmann from Helmholtz-Zentrum noted, "Our method provides an unbiased way to recognize variation in footprints and test hypotheses about their makers. It's an excellent tool for research, education, and even fieldwork."
Professor Steve Brusatte from the University of Edinburgh added, "This study is an exciting contribution for paleontology and an objective, data-driven way to classify dinosaur footprints—something that has stumped experts for over a century. It opens up exciting new possibilities for understanding how these incredible animals lived and moved, and when major groups like birds first evolved. This computer network might have identified the world's oldest birds, which I think is a fantastic and fruitful use for AI."
By democratizing analysis, DinoTracker invites amateurs to contribute to fossil studies, broadening access to paleontology.