What if birds walked the Earth tens of millions of years earlier than we thought? This mind-boggling idea is at the heart of a new controversy in paleontology, sparked by an AI system that’s turning fossil analysis on its head. A team of European researchers has developed a groundbreaking tool that’s uncovering dinosaur tracks eerily similar to those of modern birds—from a time when birds shouldn’t have existed. But here’s where it gets controversial: these footprints date back to the Triassic and early Jurassic periods, long before the earliest known bird fossils. Could this rewrite the history of avian evolution, or is there more to the story than meets the eye?
The AI, detailed in a January 2026 study published in Proceedings of the National Academy of Sciences, uses unsupervised machine learning to analyze fossilized footprints. Unlike traditional methods that rely on human classification, this system identifies patterns directly from raw shape data, grouping tracks based on eight morphological variables like toe spread, digit orientation, and ground load distribution. Trained on 2,000 tridactyl (three-toed) dinosaur footprints, the model creates an ‘eight-dimensional morphospace’—a geometric framework that categorizes tracks without relying on potentially flawed prior labels. And this is the part most people miss: the AI’s classifications align with expert interpretations up to 93% of the time, but it’s the outliers that are sparking debate.
One of the most jaw-dropping findings? A subset of tracks from the Late Triassic and Early Jurassic periods exhibit narrow toe spacing, strong central symmetry, and structures eerily reminiscent of modern birds. These prints predate the earliest known bird skeletons by roughly 60 million years. Dr. Stephen Brusatte, a co-author of the study, suggests this could mean birds have a much deeper ancestry than previously believed. But here’s the kicker: the trackmakers might not have been birds at all. They could have been non-avian theropods with convergently evolved, birdlike feet—a controversial interpretation that challenges our understanding of evolutionary timelines.
Not everyone is convinced. Dr. Jens Lallensack argues that the birdlike features could result from how dinosaur feet interacted with soft or uneven ground, cautioning that morphological resemblance doesn’t automatically confirm evolutionary lineage. And this raises a thought-provoking question: Can we trust AI to reinterpret fossil records, or are we risking oversimplification of complex evolutionary histories?
To make this discovery accessible to all, the researchers launched DinoTracker, a public app that allows anyone to upload fossil images or drawings. The app compares submissions to the AI’s morphospace, standardizing a process once dominated by subjective expert judgment. With a global shortage of ichnologists, this tool could revolutionize fossil documentation, especially in under-resourced regions. But it’s not just about dinosaurs—the team plans to expand the system to include plant impressions and invertebrate movement patterns, creating a geometry-based framework for fossil classification across multiple domains.
As DinoTracker gains traction, it’s not just academics who are taking notice. Amateur fossil hunters and enthusiasts are now part of the conversation, democratizing a field long guarded by specialists. But with great power comes great responsibility. Is AI the key to unlocking Earth’s ancient secrets, or are we stepping into uncharted territory? Let us know what you think in the comments—this debate is far from over.
