As a pet parent, there is constant worry about how your pet is doing.

I’ve experienced this myself when my new cat followed me around the house, but when I approached to pick her up, she backed away.

I was puzzled by this bizarre behavior and wished I had the power to know what she really wanted.

Well, it looks like understanding what your pet is thinking is becoming more of a reality with Google’s recently disclosed study findings on this very topic.

Let’s take a look at a short summary of the research.

Research Description

According to the research, the popularity of keeping pets is rising as one-person households increase. However, this brings new challenges, animal parents can’t always stay near their pets.

This means that you have no choice but to leave them by themselves which can lead to depression. As a countermeasure, many individuals opt to leave the TV on to provide their pets some comfort. 

cat watching television

But, note that the channel or content your animal consumes is one-sided and not necessarily according to their taste, preference, or desire. Typical to humans, animals have their preferences, and it can be challenging for any animal parent to understand.

This is why researchers have innovated a communication program that allows you to understand what your pet wants.

Moreover, it can allow the animal to consume content according to its desires and preferences. Pretty amazing, right?

ALSO CHECK: Tile Mat Customizable for Your Cat Litter Box

Google Study Report

The communication system follows four units, as illustrated in the image below.

source: Google research  [1]

The figures represent:

  1. 100 is the collecting device
  2. 200 represents a server unit
  3. 300 represents a first output device
  4. 400 is the second output device

How the System Works

Now to the technical stuff, a little dry, but very cool when you understand what it is doing.

The collecting device, 100, handles the collection of the animal’s information. This information includes the animal type, actions, noises, and surroundings.

Moreover, note the actions and noises refer to the state of the current situation. The surroundings depend on whether the animal is outdoors or indoors at a particular time, and the device can be a camera module, a microphone, or both.

The server unit, 200, refers to a receiving component that takes information from the collecting device. It manages the sending and receiving of messages to and from the user.

To manage this operation, the server has a receiving unit (210), an analysis unit (220), and a search unit (230). Unit 210 receives data from the collecting device (100), and the analysis unit (220) analyzes the information.

It then projects a message to the first output device (300) for the user to understand. The analysis unit then analyses the response received from the first output device and grasps the user’s intention.

The analysis unit uses a deep learning (AI) method to understand the animal type, noises, actions, and surroundings. This is what enables it to provide an accurate analysis and message of what the specific pet wants or desires.

The AI setting comprises three layers, namely:

  1. A locally-connected layer.
  2. A convolutional pooling layer.
  3. A fully-connected layer.

In a nutshell, these three layers work hand in hand through a specified connection. The locally-connected layer or convolutional pooling layer manages the extraction of image features. 

The fully connected layer is then tasked with determining the image features correlation for a deeper understanding. That said, the locally-connected layer interconnects with the fully-connected layer whilst also connecting to the convolution pooling layer.

The search unit, 230, is tasked with searching for the results of the response message. It does this through searching for any matching content on the database or on the internet. 

This search unit produces results like moving illustrations common to an internet video platform, YouTube. However, the results are not limited to the YouTube platform as the search unit can browse for video and image results across the internet.

As the first output device, 300, manages the sending and receiving of the messages to users. It can output audio or image illustrations. In a nutshell, this first output device can be wearable, for example, an Apple Watch, which the user wears, or wearable on the pet’s leash.

After a message is received by the first output device, 300, the user can give a response. This response is relayed back to the server unit (200) for analysis and can take an audio or manual text format.

The first output device can also be or not be connected to a second output device (400). The second output device outputs content like images and videos as desired by the subject pet.

So basically, the unit is understanding what your pet likes better and then shows it to them. I know, a little hard to wrap your head around, but WOW. How amazing would it be to know that your cat likes watching bird videos, for example.


As technological advancements sweep across the globe, the art of understanding our pets becomes easier. And as you see above, animal experts are striving to make this a reality.

I can’t wait for this to progress and have a better understanding of what my pets like. Not only will it make them happier, but it will also improve our bonds.

Therefore as a pet parent, you can brace yourself for immense technological changes in the future. Moreover, what else can beat the idea of understanding what your pet desires?

cat watching cartoons on the mobile phone


1. Lee JH. Method for providing content using communication with animals and program and system therefor [Internet]. [cited 2022 Jul 21]. Available from:

Olfa T
Olfa T

Olfa knows how to get things done and has a keen business sense that others admire. She’s always on the go, coming up with new ideas! Her ability to anticipate the needs of her readers and deliver information that they want is what makes CatVills such a success. She loves cuddling her cat Picaciu. He is her inspiration.