Since the first dogs began to coexist with humans about 15,000 years ago, there is a question that we have never stopped asking ourselves: what are they trying to tell us? There are several applications and devices that promise to give us keys to understand them. But now we would have taken a giant step.
A Chinese startup has developed a device called PettiChat that presents itself as the first real-time pet translator. Simply place it on a dog or cat’s collar so that an artificial intelligence analyzes their barks, meows and body movements and converts them, just a second later, into messages that appear on the owner’s mobile phone. The idea seems straight out of a science fiction movie, but reality is much more interesting: instead of translating barks or meows, it translates patterns.
When we hear barking we usually think that the dog is “talking.” In reality, ethologists have been explaining for decades that dogs and cats do not have a symbolic language comparable to humans. Their sounds are part of a much broader communication system in which they intervene body posture, direction of gaze, position of the tail, movement of the earsthe speed of walking or even the context in which the vocalization occurs. And that is precisely what PettiChat tries to take advantage of.
The device incorporates microphones and motion sensors capable of simultaneously recording the animal’s voice and behavior. An artificial intelligence model then compares that data with millions of examples used during its training to estimate the most likely emotional state or intention of the dog or cat. Finally, transform that interpretation into simple phrases for the owner. And it does so in English, Spanish, Chinese, French, Portuguese, Korean, Japanese, Hindi, Russian and German.
In other words, you’re not translating a language, you’re interpreting a behavior. For years, so-called “dog translators” were just electronic toys that associated certain barks with random phrases. The difference is that today’s artificial intelligence can find extremely complex patterns in enormous amounts of information.
AI already identifies tumors invisible to many radiologists, recognizes diseases by analyzing the voice, or detects heart problems by listening to just a few seconds of breathing. Something similar happens with pets: the machine does not understand what a dog “thinks”, but it can discover statistical relationships between certain soundsmovements and situations that would be almost impossible for a human being to perceive.
So… does it really work? Here it is convenient to separate technology from marketing. The company claims that its system achieves more than 90% accurate and has been trained with more than a million records of animal behavior. However, these figures come from internal testing and there are still no independent, peer-reviewed studies to confirm this accuracy in real-world conditions.
Translated: it is still early to know to what extent their interpretations are reliable. In fact, even those who closely follow this type of technology insist that it would be more correct to speak of a classifier of emotions and intentions than of an authentic animal translator. The idea of communicating with other species is not new. For decades, researchers have tried to decipher the complex songs of whales, the whistles of dolphins, the calls of primates or the dances of bees.
In recent years, artificial intelligence has begun to accelerate that work. Models capable of analyzing millions of sounds make it possible to find repeated structures that previously went completely unnoticed. Some projects even They try to build “dictionaries” of sperm whale language or identify basic units of communication in different species.
PettiChat is part of that same technological revolution, although applied to the animals we live with every day. Probably no collar will ever make a dog literally say: “Today I would prefer to leave ten minutes later” either “I don’t like that thought at all”but Yes, it could help us detect situations of anxiety, pain, stress or excitement sooner.recognize behavioral changes and even improve everyday communication between people and pets. And maybe that’s the real goal. Not adapting their language to ours, but using AI to understand their different ways of communicating and that is a huge change of perspective.