Google is moving into the AI agents era with the launch of Gemini 2.0, its next-gen AI chatbot, and a limited rollout of Project Astra, a computer vision-powered AI agent capable of analyzing the surrounding environment.
Revealed at Google I/O earlier this year, Project Astra signifies a significant advancement from Google's DeepMind team. Using a smartphone camera or smart glasses, Astra can provide answers by recognizing objects and locations, like identifying bus routes or apartment codes. Improvements in Astra entail enhanced dialogue abilities, multi-language support, integration with Google Lens and Maps, memory advancements of up to 10 minutes, and reduced response latency.
Initially available to trusted testers, a public release date has not been disclosed. However, its visual capabilities evoke privacy concerns, prompting Google to collaborate with its Responsibility and Safety Committee to mitigate risks and allow users to delete sensitive session data.
Today we announced Gemini 2.0, our most capable AI model yet. With new advances in multimodality — like native image and audio output — and native tool use, it will enable us to build new AI agents that bring us closer to our vision of a universal assistant. pic.twitter.com/SuS2OZOSzT
— Google (@Google) December 11, 2024
This announcement arrives amid a slowdown in AI developments and investment, with competition intensifying between Google and OpenAI, backed by Microsoft, for dominance in the rapidly growing AI market, projected to reach $1.8 trillion by 2030.
Google has made exciting announcements, including updates on Gemini and the introduction of Project Mariner. Gemini 2.0 is set to enhance its advanced reasoning capabilities, offering improved responses for general inquiries, coding, and math, alongside faster performance compared to earlier versions.
Developers will initially get access to an experimental version, while users can try the lighter Gemini 2.0 Flash via the app starting Wednesday. This upgrade is expected to be integrated into more Google products by early next year.
Additionally, Google revealed Project Mariner, a prototype Chrome extension designed for complex task completion. It can analyze text, images, graphs, and other web elements at the pixel level, though currently limited to trusted testers, it faces challenges with accuracy and speed.
Moreover, Google is exploring AI applications in gaming, collaborating with Supercell, the creator of Clash of Clans and Brawl Stars, to develop an AI that can assist players in game-related queries. However, details regarding the AI's data sourcing remain unclear, raising questions about whether it's based on developer information or online gaming guides.