AI could soon speak dog and cat

Imagine what it would be like to know exactly what your dog was saying when it barked, or your cat when it miaowed, or your iguana when it … made whatever sound an iguana makes (do they even make one?).

There have been various attempts to develop systems that enable pets to communicate — a notable example is where dogs press programmable sound buttons, with each one representing a word or phrase, to “talk” and express needs, emotions, or observations — but none are quite so exciting as the one proposed by Chinese tech giant Baidu in a patent filed with the China National Intellectual Property Administration.

The patent describes a system that gathers animal data such as vocal sounds, behavior, and physiological signals, Sky News reported this week. The collected information would be processed and then merged for AI analysis, creating a technology capable of interpreting whatever sounds the animal makes.

But don’t expect that you’re going to start having conversations with Fido about the weather, or that your cat will start making judgmental remarks about your Netflix choices. For a start, there’s no mention of creating an AI that can speak back to animals, and the main information you’re likely to get is something about your pet’s emotional state.

In the patent, Baidu says the system would improve “the accuracy and efficiency of cross-species communication,” though a spokesperson for the company cautioned that its technology is “still in the research stage.”

There has already been some high-profile research into trying to find ways to accurately translate animal noises, particularly those of dogs. An ongoing study by UC San Diego’s Comparative Cognition Lab, for example, has been looking at how the previously mentioned soundboards can be used to help dogs express themselves. Some dogs have even shown that they can combine multiple words to form simple requests like “outside play” or “food now.” But critics argue that there may be too many assumptions about the dogs’ ability to fully understand the words, suggesting that they could just be pressing the buttons because they’ve learned to do so in response to certain cues, like the button’s sound.

With AI tech advancing so rapidly, let’s hope Baidu is onto something here.