by In the Dr. Dolittle books and movies, the ability to “talk to the animals” captured the imagination. Now a $10m prize is being offered to scientists to create real conversations.
The Jeremy Coller Foundation and Tel Aviv University launched the Coller Dolittle Challenge for Two-Way Interspecies Communication. Although the use of AI is not mandatory, the team says the technology can boost almost every proposal.
“Just as the Rosetta Stone unlocked the secrets of hieroglyphics, I am convinced that the power of AI can help us unlock interspecies conversation,” said Jeremy Coller, chairman of the foundation.
The team says the idea has precedence: researchers have recently developed machine-learning algorithms to translate bat squeaks, allowing them to identify the content of the squabbles.
Other efforts in the field have included algorithms to decode the emotions of pigs from their grunts, and rodent squeaks to recognize when they are stressed. The Earth Species Project, a California non-profit group, is also working in the area, hoping to develop a system that can be applied to all species.
“In recent years, there has been significant progress in the scientific community’s understanding of the communication patterns of non-human organisms,” said Professor Yossi Yovel of Tel Aviv University, who chairs the Coller Dolittle Prize and co-authored the study bat. .
Related: Can artificial intelligence help us talk to the animals?
While the grand prize for “cracking the code” is an equity investment of $10m (about £7.8m) or a cash prize of $500,000, there will be an annual prize of $100,000 to help researchers in the field come up with ” models and algorithms scientifically rigorous for coherent communication with non-human organisms until interspecies communication is achieved”.
Criteria for the small prizes include that approaches are non-invasive and applicable to a range of contexts, are based on normal animal communication signals, and allow researchers to respond to the animal’s attempts to communicate with it to measure.
“We hope to announce the criteria for the grand prize after two or three years of small prizes,” said Yovel.
The team behind the prize, which closes for submissions on July 31, said the goal is to develop a system where animals don’t realize they’re actually communicating with humans – like the Turing test for AI, where it’s like goal of creating a system. a computer system whose conversation with a person is indistinguishable from that of a real person.
“We are open to any organism and any modality from acoustic communication in whales to chemical communication in worms,” said Yovel.
The team says the prize could have important consequences for understanding animal sentience, and therefore support for animal rights.
Peter Gabriel, musician and co-founder of Interspecies Internet, who helped develop the concept of the prize, said: “When I was playing music with the bonobo apes, I was blown away by their intelligence and musicality … I’m delighted to be there. are scientists now serious about understanding their communication and ways by which we could begin meaningful interspecies communication.”
Dr Katherine Herborn, an animal behavior researcher at the University of Plymouth, said that one potential benefit of technologies to decode animal communication was to gain an understanding of farm animals that might be needed to improve their management.
However, some experts have raised ethical questions about conversing with animals and questioned whether AI can really shed light on the meaning or function of animal vocalizations.
“I think no amount of AI programming can replace long-term, detailed knowledge of the society in which animals communicate,” said Robert Seyfarth,
professor emeritus of psychology at the University of Pennsylvania.
“Trying to find out the meaning of a baboon’s grunt, a dolphin’s whistle or an elephant’s roar without knowing the social background – this takes years – is like jumping to page 137 of Pride and Prejudice and trying to explain why does elizabeth bennett say no to mr darcy. without any knowledge of its history.”
Clara Mancini, professor of animal-computer interaction at the Open University, said it was very plausible that AI would help us understand animal communication. “If successful, in my opinion, this would be one of the most remarkable human achievements of this rapidly developing technology,” she said.
But, she said, whether our success in this great challenge will really allow us to understand that the experience of animals is a different matter.
“More importantly, however, is the question of whether we will be willing to truly listen to what animals have to say and finally give them the basic rights that their intrinsic dignity demands. I sincerely hope.”
