Google announced the new version of its multi-modal artificial intelligence system Project Astra at the I/O 2025 event. Able to perceive its surroundings in real time and answer questions about the environment in a chat form, Astra will now have a much more useful structure.

Google announced the new version of Project Astra

The new version is equipped with advanced features such as longer-term memory, control capabilities on the computer, and the ability to complete multi-step tasks. This version can perform much more comprehensive tasks by remembering the user’s previous commands and interactions.

In the promotional video released by Google; Astra is seen searching the internet to repair a mountain bike, gathering the necessary information, reviewing relevant e-mails, and even calling a local bike shop for a spare part.

These capabilities will be one of the concrete steps of Google’s goal of turning the Gemini artificial intelligence platform into a “universal assistant.” DeepMind CEO Demis Hassabis said that this model is not just a system that answers questions; It also states that it is an artificial intelligence tool that takes on daily tasks, tracks work on behalf of the user and increases productivity.

Astra’s multimodal structure includes the ability to recognize visual data, process natural language and integrate these two areas to perform tasks in real-world scenarios. Some of these capabilities are already included in Gemini Live. However, according to Google’s statements, Astra’s ultimate potential will be revealed when these technologies are transferred to devices such as smart glasses.

So what do you think about this? You can share your views with us in the comments section below.