At Google I/O 2024, the search engine presented a new virtual assistant, Project Astra, based on the Gemini language model, capable of analyzing video, voice and text to respond to all questions. questions.
Google Assistant’s hours now seem numbered. Google has just announced, during its Google I/O conference dedicated to its software innovations, a new assistant named, for the moment, Project Astra. The latter is based, unsurprisingly, on the language model of the American brand, Gemini. The latter uses Google’s computing power to analyze text, voice and image to obtain contextual answers to each question asked.
An assistant who always listens and sees what you see
Project Astra was not presented on stage, but through a video, filmed in one go and without editing, according to the Google teams. This technology is currently not a finished product, but a working project from Google DeepMind, the team responsible for research in artificial intelligence.
In this excerpt, we can see a user launching the voice assistant, then opening her smartphone camera. This way, Project Astra can see what the demonstrator sees to provide answers to all her questions.
In the examples presented, Project Astra was able to understand a piece of code filmed by the camera and give indications for improving it. It can also recognize objects, or give suggestions based on the items in front of it and the questions asked by the user.
A research project which foreshadows Google’s ambitions in the coming years
Even stronger: Project Astra analyzes a number of data when the phone is moved from one place to another, and in real time. In this same video, the user asks where her glasses are, and the artificial intelligence is able to remind her of the exact location where they were left.
Project Astra is only a proof of concept and won’t make it to Android for probably several months, if not years. It is also very close to GPT-4o, the latest OpenAI language model presented this Monday, May 13, and which has the same functionalities, with a voice that could be described as more natural.
Before that, users will be able to benefit from Google’s advances in the field of artificial intelligence with the integration of Gemini into all Google services, particularly on the search engine, in Gmail or in Google Workspace office applications.
Source: Google I/O Conference
3