From Liam Mo and Brenda Goh at Reuters, we learn that China’s largest legal search engine is trying to patent “ a system to convert animal vocalisations into human language”:

The document says the system will collect animal data, including vocal sounds, behavioural patterns, and physiological signals, which will be preprocessed and merged before an AI-powered analysis designed to recognise the animal’s emotional state.

The emotional states would then be mapped to semantic meanings and translated into human language.

The system could allow “deeper emotional communication and understanding between animals and humans, improving the accuracy and efficiency of cross-species communication,” Baidu said in the patent document.

“China’s Baidu looks to patent AI system to decipher animal sounds,” May 9, 2025

The problem we face

In most cases when we think we have deciphered a message, from a cat, for example, what we really have is a human interpretation of what the cat is expressing. How much difference does that make?

Let’s say Tabby is sitting in the front window. He sees another cat crossing the yard, He stiffens and growls, hair standing on end.

The new AI helpfully translates: “I do not like that cat.”

Tabby scurries to the door and yowls. The AI helpfully translates: “Let me out right now!”

The human, who works from home, decides that breaking up a catfight is not on the agenda for this afternoon. So he does not open the door. He puts Tabby in the basement instead. Tabby continues to yowl from behind the basement door.

The new AI helpfully translates: “Let me out right now!” “Let me out right now!”

It probably won’t be the same experience as an AI translating your new neighbor’s complex refugee story from the Uyghur language…

I’m not saying the system couldn’t be helpful. But if we assume that cats have a language that requires translation the way human languages do, we are most likely to find instead that the cat’s behavior is his language. Caution: The rest may end up being human imagination.