Does artificial intelligence understand animal language?
Maybe you’re wondering what animals say to each other. Now, thanks to artificial intelligence, you can get answers to many of your questions.
Ghazal Ziari: Have you ever asked the question of what the birds are talking about?
You may want to know what cats think of you.
Yossi Yovel and Oded Rechavi from Sagol University’s Faculty of Zoology, Biological Sciences, and Neuroscience studied the understanding of animal communication using artificial intelligence.
“Artificial Intelligence and Dr. In the article titled “Doolittle’s Challenge,” the team examines the potential and limitations of animal intelligence, communication, and artificial intelligence capabilities in this field.
“Dr. They described it as the “Doolittle Challenge.” In this fictional story, Dr. Doolittle learns the language of birds from a Polynesian multilingual African parrot who speaks both avian and human English.
He says the animals speak another language in this Polynesian tale: “You’ll understand if I say Polly wants a piece of biscuit, but listen to this: ‘Kaka-oo-yeee, fi-fi.’”
In response, the doctor says, “Well, What does that mean?”
And the Polynesian’s answer is, “Does that mean, in the language of the birds, that the porridge is still hot?”
Regardless of the abilities of animals, humans, or artificial intelligence, the text and context of communication is very different from directly assigning meaning to sounds, according to the researchers, just like different human languages. Various communication factors are needed to communicate with the animal kingdom, and scientists have identified three major obstacles in this direction.
Background
The first limitation is the difficulty distinguishing the context of animal communication. Humans have been recording animals’ communication and imitating their voices for decades, and researchers can get responses from different animals based on this technique without the help of artificial intelligence.
For example, a half-female robot frog attracts real male frogs to mate, and a fish robot interacts with live fish, influencing their movements.
A bee robot also managed to get real bees to dance and fly to a specific place in the field. This particular dance is actually a way for the bees to convey information about the location of the food and to show each other a certain direction with certain movements.
These examples demonstrate the potential of engineered stimuli to elicit responses and even transmit information to animals in context-specific situations.
Where AI can produce sounds similar to animal communication sounds, it will be difficult to determine whether they approximate a context beyond familiar sounds or animal-like sounds.
Once trained with recorded sounds, the AI can simulate the sound of a particular bird’s call but determine whether that bird makes that sound to mark its territory, to attract a mate, or
sing or convey another message. It needs more information. animal language
The authors of this article believe that this story also applies to silent communication and insects that communicate using chemical signals. Without observing the insects’ behavior, it will be difficult to decipher whether the chemical is a mating signal or a warning of danger.
AI needs human input and data to start working on animal communication, including bird songs. Human biases strongly influence the interpretation of these signals. Sometimes a bird’s song needs to be set as the background for artificial intelligence so that it can be linked to the songs of other birds. This requires a controlled approach to observing animal communication that understands only its natural responses. animal language
Inference of natural reactions
Animals display a wide variety of behaviors, and a variety of factors, such as the current physiological state, social dynamics, and environmental conditions, influence their responses.
Different animal species use different sensory methods to communicate, such as sound, chemical signals, or body language, and discerning specific communication-related reactions may require a range of observational techniques. Experiments cannot train animals to perform reactions, a common research technique, because this method closes the window of natural communication and makes laboratory observations dependent on training.
Measuring responses can also be challenging Because there may not be a clear and measurable reaction from the outside. People may miss some reactions. AI algorithms are trained to interpret responses; this increases the likelihood of seeing subtle movements in the responder’s eyes and associating them with something that isn’t really there.
The third barrier identified involves animal communication, which focuses on a limited set of contexts. These limitations potentially limit the range of communication between species, reducing the ability to communicate with animals in a wide variety of topics or contexts. animal language
potential benefits
Of course, there are many potential benefits associated with animals. Effective communication with honey bees or other pollinating animals can benefit agriculture. Recognizing the moods and moods of farm animals will help them take better care of them, and understanding the whispers of animals in a protected forest will alert conservationists to an unknown threat. Imagine that the process of research experiments would have been significantly improved if the mice could share their views.
It also serves as a first contact app for potential extraterrestrial encounters. If we cannot communicate with sentient beings and intelligent life on our own planet, what hope do we have of communicating with possible beings on other planets?
