Project Astra: Google AI that understands the environment

During the event Google I/Owhich took place this Tuesday (14) in the United States, the Search giant revealed to the world its newest tool based on artificial intelligence. The tool in question is capable of using the smartphone’s camera to analyze and interpret the environment around us.

Read too
How to disable Google Assistant: A step-by-step guide
You can now download the Gemini app for Android in Brazil. Know how!

Astra Project

Currently known as Astra Projectit stores in memory information about what it visually perceives, thus making it possible to provide increasingly assertive responses to its users.

The company’s executives emphasized that what was presented at the event represents just a preview of what is being developed internally in its production lines. They also suggested that, as the project evolves, it is expected that several of its innovative features will gradually be integrated into other already established products and services in the Google ecosystem.

Even though it was still at an early stage of development, Project Astra managed to impress and draw applause from those present at the convention center in Mountain View.

The system designed by Google is versatile, operating not only through smartphones but also through augmented reality glasses, the specific technical details of which have not yet been released to the public.

Demo video impresses

A demonstration video was shown where a woman asked Astra to locate an object in the room that was capable of reproducing sounds. With remarkable efficiency, the system promptly identified and highlighted an available speaker in the environment.

Then, the user, moving towards the speaker, drew an arrow on the screen of their mobile device and asked a question about which component the arrow was pointing to. With precision, Project Astra recognized the component and reported that it was a tweeter, showing its ability to understand things around it.

Not stopping there, Project Astra demonstrated its ability to perform logical reasoning or pranks. When confronted with a drawing containing two cats — one alive and the other dead — both situated next to a representation of a box, he astutely conjectured that it was the famous mental experiment known as Schrödinger’s cat.

In an additional scene, when a participant pointed his cell phone towards the window and asked about its location, the intelligent system did not hesitate to recognize and announce the name of a famous London neighborhood. The disclaimer at the beginning of the video says that everything was recorded in just one take. Really impressive.