A group of researchers has won a $100,000 prize aimed at enhancing communication with animals by uncovering the meanings behind dolphin whistles.
The Collad Rittle Award, established last year by the Jeremy Coller Foundation and Tel Aviv University, focuses on two-way communication.
The team from the Sarasota Dolphin Research Program, including Laera Sai and Peter Tayak from Woods Hole, has spent over 40 years studying bottlenose dolphins in Sarasota, Florida.
The researchers employed non-invasive methods, such as attaching a hydrophone to a suction cup and using a digital acoustic tag to capture dolphin sounds, including what are referred to as “signature” whistles, akin to names, and “nonsignature” whistles, which are largely misunderstood despite making up about 50% of dolphin calls.
In their latest study, though it’s not yet peer-reviewed, the team found at least 20 types of nonsignature whistles produced by various dolphins, noting that these two categories are recognized by at least 25 individuals.
When researchers replayed these sounds to the dolphins, they observed evasive behaviors, hinting that one sound could act as an alarm while the other may signal an unexpected encounter.
Sayigh expressed surprise at their win, stating, “I’m not excited because I really didn’t expect it. It’s a great honor.”
The review panel was chaired by Yoshijobel, a zoology professor at Tel Aviv University. Interestingly, his team had previously avoided using machine learning to decode animal sounds.
“We were quite impressed by the extensive long-term dataset created, and we believe this will lead to more significant findings,” said Yovel, appreciating the team’s non-invasive recording techniques and their innovative field studies involving drones and speakers.
Jubel emphasized that this achievement might pave the way for applying AI to data, potentially yielding even more fascinating results.
Jonathan Burch, a philosopher from the London School of Economics and a judge, highlighted the scarcity of data as a major hurdle in deciphering animal communication.
“Consider the massive amounts of data required to train a large language model like ChatGPT. There’s nothing comparable in other animals,” he remarked.
“That’s precisely why projects like the Sarasota Dolphin Research Program, which has been collecting an extensive library of dolphin whistles for decades, are crucial. The cumulative effect of their work allows Sayigh and her team to utilize deep learning in analyzing these whistles, perhaps ultimately leading to a breakthrough in understanding,” he added.
Yovel mentioned that about 20 teams participated in this year’s contest, with four finalists emerging, including those studying communication in Nightingales, squid, and marmots. He added that applications for the 202-26 awards are currently open.
Beyond the annual $100,000 prize, there’s also a grand prize involving either a $10 million investment or $500,000 cash. To achieve this, researchers must create algorithms enabling animals to “communicate independently without realizing they are interacting with humans.”
This challenge draws inspiration from the AI Turing test, where humans shouldn’t be able to distinguish between a system conversing with them and an actual person.
Robert Seyfers, a psychology professor at the University of Pennsylvania, commended the researchers, stating, “These scientists are pioneering a new understanding of dolphin communication and cognition, which is truly groundbreaking.”
Clara Mancini, an expert in animal-computer interaction at Open University, remarked that this dolphin research illustrates how technology can enhance our understanding of animal communication, potentially allowing future interactions on their terms.
“One of the primary benefits of such advancements is that animal communication systems can ultimately demonstrate the nuanced and effective nature of human language, shaped by our own evolutionary context,” she noted.
“However, embarking on the path to interspecies communication involves recognizing that deciphering language is different from grasping the lived experiences of language users. We must remain aware that both curiosity and the challenges ahead require us to respect the unique knowledge and perspectives of different species.”
