Now you can communicate with your pet with the animal-to-speech translator

By Aswin Anilkumar

The idea of humans communicating with animals has been a distant dream. From the early Graeco-Roman myths about demigods conversing with their pets to the popular modern movies such as Dr Doolittle, the idea of conversing with the animal kingdom has piqued human interest for centuries.

Considering how deeply involved animals are in human activities, it’s no wonder a rudimentary system of communication has evolved. Such as when dogs are taught to fetch and humans learn to distinguish between a playful and a threatening bark. Nevertheless, there exists a limit to this communication. For example, the inability of humans to distinguish between a scared animal and an aggressive one can sometimes result in a tragedy.

Armed with Artificial Intelligence (AI) and algorithms, Dr Con Slobodchikoff at the Northern Arizona University is working to change this by building a reliable human-dog translator.

Talking animals and anthropomorphism

While there have been recorded instances of ‘talking’ animals, much of their speech is mere vocalisation. That is, these animals do not understand what they speak. They merely imitate sounds and demonstrate ‘intelligence’ by observing visual cues of their trainers. Clever Hans was one such horse who could apparently count and tell the date by tapping his hoof. However, it was later found out that Hans would merely watch his trainer relax when the right number was tapped—disabusing any notion of Hans’ intelligence.

Even as a multitude of studies demonstrate that animals do not truly speak, the case of Alex the African Grey Parrot is unique. Alex was adopted when he was one year old by Dr Irene Pepperberg. Over a period of 22 years, she taught him to identify seven colours and five shapes. He could understand comparative terms and abstract concepts such as the number zero. However, some people criticise Alex’s ‘speech’ as being mere sounds produced for earning a reward. Yet the scope and breadth of the experiment to train Alex contributed to the development of various methods to teach the learning-impaired and the autistic.

A key bias in evaluating whether animal speech is truly conscious arises from our projected anthropomorphism—the attribution of human characteristics to an animal or object. We project human traits onto animal behaviour. One such incidence was when a widely shared image allegedly showed a kangaroo cradling a dying mate. However, scientists later clarified that the act was probably carnal in nature. It is this anthropomorphism which convinces us that animal speech—mere sounds as a response to a stimulus—is sentient and intelligent. As animals cannot learn human languages and respond; humans need to learn to understand animal speech and respond appropriately.

Dr Slobodchikoff’s revolutionary research

Dr Slobodchikoff has proposed using AI to decode animal expressions and sounds. His pioneering three-decade long research on prairie dogs convinced him that the animal’s elaborate vocalization was a rudimentary language. He realized vocalisations of animals like dogs and cats could also be decoded and translated. Dr Slobodchikoff’s method involves sifting through a large library of dog barks. The computer is then taught to identify patterns in, say, an ‘angry’ bark or a ‘sad’ bark. This is followed by repeated testing to refine its understanding. Eventually, the machine’s learning would become comprehensive enough and it can be translated directly into English speech.

The benefits of animal-to-speech translators are not confined to bonding between a pet and a master. Translators could help farmers tell which of its animals are in pain, improve research into animal perception, and could even help train sniffer dogs to communicate the difference in scents. The potential applications are enormous in scope. The invention of a working animal-to-speech translator would mark a significant moment in human history.

Featured Image Source: Pexels