Alphabet, the parent company of Google, presented on Tuesday at the annual developer conference Google I/O an innovative artificial intelligence capable of instantly responding to video, audio, and text queries. With this major announcement, the tech giant aims to demonstrate its leadership in the field of AI and counter criticisms about its alleged lack of progress compared to its competitors.
At the conference, Google CEO Sundar Pichai unveiled this new multimodal AI assistant, called “Project Astra”, which is powered by an enhanced version of their Gemini model.
Astra is one of the standout elements from Google’s recent wave of announcements where the company showcases its vision in the world of AI, following the recent product launches and updates from rival companies like OpenAI, Meta, and Microsoft. In an impressive demonstration, Google’s assistant prototype responded to voice commands by interpreting what it sees through a phone camera or smart glasses.
Pichai announced Google’s plans to integrate Astra’s features into the Gemini app this year, as well as into all the company’s products, although he was cautious and prioritized quality over speed in launching it. Demis Hassabis, from DeepMind, wanted to highlight the technical complexity of achieving real-time conversational responses, emphasizing advances in spatial comprehension and video processing.
Google also introduced significant changes to its search engine, with an “AI summary” at the top of many results, followed by links and ads. These changes, which aim to simplify searches and will expand worldwide this year, come at a time when OpenAI threatens Google’s dominance in search thanks to its chatbot ChatGPT, capable of providing quick and comprehensive answers, even through voice and video.
To counter the threat of OpenAI, Google has unveiled new AI products such as Veo for generating video from text, Imagen 3 for creating images, and Lyria for music. In addition, Gemini Advanced subscribers will be able to create custom chatbots called “Gems”, and the Gemini 1.5 Pro model has been updated with a larger context window.
