AI uncovers rapid evolutionary bursts in songbirds linked to climate shifts
University of Michigan researchers used AI to show that songbirds evolved in sudden bursts that line up with major climate changes.
A University of Michigan team applied artificial intelligence to digitize and measure thousands of bird skeletons, creating a dataset of more than 170,000 bone measurements from over 2,000 passerine species. Using the AI system Skelevision and a novel statistical framework named bifrost, they reconstructed roughly 45 million years of morphological evolution. The analysis identified distinct bursts of rapid change, notably around 35 million years ago during a major global cooling transition, and periods of slowdown about 15 million years ago.
The researchers further showed that species living at extreme latitudes tend to evolve faster than those near the equator, suggesting climate variability drives morphological innovation. The study, published in Nature Ecology & Evolution, underscores the untapped research value of museum collections when paired with modern AI tools. Funding came from Schmidt Sciences, the Packard Foundation, and several governmental science agencies.
Why it matters
Understanding how past climate upheavals spurred rapid evolution helps predict how current warming may reshape biodiversity.
In this story