Advancements in AI may soon allow humans to talk with animals, but some groups claim the research could be violating the animals’ right to privacy.
Machine learning models are paving the way to catalogue and even mimic the sounds animals make to communicate with each other, with researchers touting the technology as a way to better understand the world of animals.
But “one risk that is currently not taken into account in research ethics is the concern that we might be violating the privacy of animals by collecting data on their vocalizations and movements,” according to researchers in an article published in Springer Nature.
“An interest in privacy is an interest in controlling how you appear to others, and the complex social lives of many nonhuman animals offer strong support for thinking animals have this interest.”
The ability to “speak” with animals also could pose dangers to the natural world, experts said.
The controversy involves a technique known as “playback,” which is the practice of broadcasting recorded animal calls in the wild.
Biologist Vittorio Baglione, who has spent his career cataloging the calls of carrion crows, was blown away by how AI models could help sort his recordings and point out commonalities in specific cries.
Among the calls was one that Baglione described as essentially, “Hey! Come help!” as the crows would use it to summon others to help defend the nest, he told Bloomberg.
But Baglione said his research and the breakthroughs in communication could be co-opted by hunters or anyone with ill intent to lure in the crows.
“You can stress the animals. You can disrupt their social environment. You can pose conflicts between them, especially if you don’t really understand the meaning of the call,” the scientist said of such playback.
This is why playback remains controversial, as it takes advantage of an animal’s trust in a specific sound to make it act in whatever way the user wants, explained César Rodríguez-Garavito, director of the More-Than-Human Life, or MOTH, Program at New York University.
In his 2025 report on the development of AI models striving for animal communication, Rodríguez-Garavito said the technology has the capacity to expose animals to mobs of people wanting to chat or to even weaponize them.
Researchers and animal-rights groups have pointed to previous playback experiments to show the distress an animal can suffer when confronted with an imitation call from a machine.
In 2007, zoologist Katy Payne admitted she caused an elephant great distress when playing a recording of her dead grandfather’s voice.
Despite the concerns, there remains a $10 million reward for a project that can establish communication with an animal without that animal recognizing that it is communicating with humans.
Private equity billionaire Jeremy Coller, who funds the Coller Doolittle Challenge, said during the challenge’s latest prize event in June that humans will likely reach this goal within the decade.
“Because AI is so fast, I have absolute conviction we will crack the code by 2030,” he said.
