Google unveils Project Astra at I/O 2024 conference: Check features
What's the story
Google has introduced its latest artificial intelligence (AI) innovation, Project Astra, at the I/O 2024 conference. The company's DeepMind CEO, Demis Hassabis, expressed his team's ambition to "develop universal AI agents that can be helpful in everyday life." Project Astra is a significant stride toward achieving this objective. The project uses a phone's camera as its primary interface, and is designed to assist users in identifying objects in their environment.
Prowess
Project Astra has impressive capabilities
Project Astra's AI, named Gemini, demonstrated its advanced capabilities in a video released by Google. In the demonstration, a user asked Gemini to identify items that make sound. When the camera focused on a speaker next to a monitor, Gemini accurately responded with "I see a speaker, which makes sound." Furthermore, when queried about specific parts of the speaker, Gemini was able to correctly name them.
Creativity
It showcases creative response and memory recall
Gemini has also demonstrated its ability to generate creative responses. When presented with a cup of crayons and asked for a creative alliteration, Gemini responded with "Creative crayons color cheerfully. They certainly craft colorful creations." In addition to creativity, the AI showcased impressive memory recall capabilities. When asked about the location of an item not currently in view, Gemini accurately remembered and stated "Yes, I do," adding "Your glasses were on a desk near a red apple."
Efficiency
Real-time data processing and problem-solving
Project Astra demonstrated its potential for real-time data processing and problem-solving. In one instance, while viewing a diagram on a whiteboard through the camera of wearable glasses, Gemini suggested ways to improve system speed when asked by the user. "Adding a cache between the server and database could improve speed," it advised. The AI's ability to make connections and associations was also highlighted, when it identified a doodle of two cats on a whiteboard as "Schrodinger's cat."
AI development
Project Astra's AI enhancements and future prospects
Hassabis revealed that capabilities of Project Astra are possible because the agents were "designed to process information faster by continuously encoding video frames, combining the video and speech input into a timeline of events, and caching this information for efficient recall." Google is enhancing its AI's vocal expression range to provide a wider range of intonations. While Project Astra has no confirmed launch plans, Hassabis stated that these assistants could be available "through your phone or glasses" in the future.