AI for Animals 08.14.17

Artificial Intelligence Translation Device for Animals Within a Decade

Paging Dr. Doolittle: Artificial intelligence translation technology capable of interpreting dog language could become a reality within 10 years, according to a report from futurologist William Higham working in cooperation with Amazon.

“Innovative products that succeed are based on a genuine and major consumer need,” Higham commented to The Guardian. “The amount of money now spent on pets—they are becoming fur babies to so many people—means there is huge consumer demand for this. Somebody is going to put this together.”

Higham cited research by Con Slobodchikoff, professor emeritus at the department of biological sciences at Northern Arizona University, on animal communication. Slobodchikoff has spent a generation analyzing the behavior of prairie dogs, leading him to determine the animals have “a sophisticated communication system that has all the aspects of language.”

“They have words for different species of predator and can describe the color of clothes of a human, or the coat of coyotes or dogs.,” Slobodchikoff said.

Slobodchikoff told CBC News he has been collaborating with a computer scientist to develop a system that uses voice pattern recognition techniques and artificial intelligence translation to interpret animal speech.

“We could potentially have something maybe the size of a cellphone in five to 10 years where a dog would say, ‘Woof’ and the device would say. ‘I want to eat chicken tonight” or a cat could say, ‘Meow,’ and the device would say, ‘My litter box is filthy, please clean it.'”

Taking on animal speech represents an entirely new horizon for artificial intelligence translation technology. Systems such as Google Translate have made major strides in human translation in recent years, with the Google technology using neural machine translation technology set to support interpretation for 103 languages.

New earpieces developed by companies including Lingmo International and Waverly Labs are able to instantly translate various human languages.

Advances in the field of machine learning have led to improved speech recognition technology. New algorithms can learn to interpret languages on their own by analyzing massive sets of data.

However, it’s unclear whether such technology can really deliver accurate interpretation of animal communications.

“We would not describe dogs’ forms of communication as language in the scientific sense,” said Juliane Kaminski, a psychologist at Portsmouth University, in comments to The Guardian. “They do give out rudimentary signals of what they want and how they’re feeling.”

Because of this, animal language may not be decipherable by AI translation technologies.

Nirel Marofsky is project analyst for the cognitive engine and application ecosystem at Veritone. She acts as a liaison to strategic partners, integrating developers and their capabilities into the Veritone Platform. Learn more about our platform and join the Veritone developer ecosystem today.