A University of California, Berkeley researcher who spent more than a decade listening to zebra finches has won a prize for two-way communication with another species, she told Live Science. Julie Elie, an ethologist and associate project scientist at Berkeley, recently won the 2026 Coller Dolittle Challenge for Two-Way Interspecies Communication for work that combines thousands of recordings, machine learning, and behavioral experiments. In an interview with Live Science's Kenna Hughes-Castleberry, she said bird calls are more than pleasant sounds. They can reveal how birds think, learn, cooperate, and respond to their surroundings.
Elie came to birds somewhat by accident. As a young person she was more drawn to mammals, she said, and she started her Ph.D. interested in how animals interact and what social organization they establish. A mentor suggested the zebra finch. In the wild, she said, it is hard to watch birds closely because they fly off. In the lab, she could watch a colony in cages where the birds were fully interacting. The most important relationship, she found, was not a dominance hierarchy, as some earlier work had assumed, but the pair bond. Males and females pair for life, and that partnership is the core of their social organization.
Zebra finches learn vocalizations in a way that has made them a classic model for song and speech. Elie said researchers have long wanted to understand the neural mechanisms that let a bird listen to a sound, copy it, and reproduce it. By recording thousands of calls and pairing those sounds with what the birds were doing, she and her colleagues have used machine learning to pick out specific calls and test what they mean in behavioral experiments. The prize, named for the Coller Dolittle Challenge, is intended to push researchers toward genuine two-way communication rather than one-way decoding. Elie told Live Science that the first species people might be able to communicate with vocally might be a bird.
The finding is still a research program, not a finished translation app. Elie's work shows that some calls can be classified and linked to social context, especially around pair bonds, and that AI can help sort a large library of sounds that a human listener would struggle to keep straight. It does not mean scientists can hold a free-form conversation with a finch. What it does mean, in her account, is that a careful mix of observation and computation can turn a supposedly simple cage bird into a window on animal communication, and that Berkeley's long-running zebra finch work now has a high-profile prize attached to that claim.