When researchers fail to find complex communication in an animal, or can't work out what its signals mean, they tend to conclude it isn't there. The animal gets blamed for failing a test that really measured the limits of the human observer.
Machine learning can translate between human languages without a bilingual dictionary. It does this by finding deep patterns in how each language is structured that people could not perceive. The same approach might let us decode an animal's communication with no shared key, which turns 'talking to animals' from fantasy into a technical problem.
Automated recording and AI let biologists capture what used to slip away: movements, interactions and calls of individual animals, tracked across years and oceans. Biology is shifting from mostly studying bodies to studying behavior and communication, and what used to be anecdote is now a data point.
The closer we look at other animals, the more traits once thought uniquely human turn up in them. Whales, for example, form clans defined by how they sound, and they teach learned behaviors that pass down as culture. Each finding weakens the assumption that humans are fundamentally special.
People mainly protect creatures they can empathize with. Learning that whales use their voices in ways like ours set off a wave of empathy that helped end their slaughter. Understanding animals' signals could deepen that change, but the same knowledge brings a power to manipulate and exploit that has to be guarded.