Google IO 2025 Updates

Gemini Live brings AI to your camera, enabling real-time object recognition and contextual understanding.
Point your camera at anything and ask: "What is this?" The AI instantly identifies the object.
Further, ask: "How can I use this?" or "What can I create based on this?" for ideation assistance.
Point your camera at text on paper, ask questions and receive ideation support.

Gemini Live brings AI to your camera, enabling real-time object recognition and contextual understanding.
Point your camera at anything and ask: "What is this?" The AI instantly identifies the object.
Further, ask: "How

Vevo 3 enables users to create movies from text prompts.
Features include:
- Native sound integration
- Cinematic quality visuals
- Integrated music
- Camera movement simulation
Example: A text prompt like "A wise old owl and a nervous young badger in the forest" can generate a complete movie scene.

Project Astra is a prototype, analogous to Jarvis from Iron Man, designed to be a comprehensive AI assistant.
Functionality:
- Observing and understanding your environment.
- Assisting with various tasks.
- Operating other systems.
- Providing support for work, projects, and explanations.

Google Flow is a tool for creators that converts text into film.
Process:
- Input text.
- AI converts text to script.
- AI creates transitions and visual text.
- Generates a complete movie.
Simplifies short film creation by automating video generation from textual scripts.
Example scenario: A content creator inputs points about Google IO 2025 updates, and the AI generates a complete video.

Agent Mode acts as an intelligent agent that executes tasks based on user input.
Functionality:
- Takes spoken commands and executes them.
- Example: Planning a trip to Goa for six days within a budget of $50,000, including ticket booking.
- Completes tasks end-to-end, requiring only user confirmation and payment approval.
- Fills out online forms and handles various tasks beyond travel planning.
Gemini app with Agent Mode can find apartment listings meeting specific criteria (location, roommates, budget) and schedule tours.
Agentic capabilities integrated into Google Chrome, Google Search, and the Gemini app enable these platforms to act as agents for task completion, not just search and information retrieval.

Google Jules is an AI tool designed to assist with coding tasks.
Capabilities:
- Analyzes existing code.
- Edits and debugs code.
- Writes new code.
Designed for both senior and junior coders to enhance productivity and code quality.

Google Meet AI Speech Translator provides live translation during video calls.
Functionality:
- Translates spoken language in real-time.
- Example: A user speaks in Hindi, and the recipient hears the translation in Japanese.
Facilitates global communication by eliminating language barriers.
Example Scenario: Booking a vacation rental in South America with real-time translation from the property owner.

Google Beam converts 2D images from video calls (e.g., Google Meet) into 3D experiences.
Features:
- Creates the sensation of being in the same room as the other person.
- Uses multiple cameras to capture different angles and AI to merge video streams.
- Renders a 3D light field display with precise head tracking and real-time processing at 60 frames per second.
  $(6 <br>umber of cameras)$
  $(60 frames per second)$

Google Try On allows users to virtually try on clothing online.
Process:
- Upload a photo.
- Select clothing items.
- Virtually see how the clothes look on the user's photo.
Helps users make better online shopping decisions.

Project Mariner is a research prototype AI agent designed to interact with the internet and perform tasks.
Key Features:
- Multitasking: Handles multiple tasks simultaneously.
- Learning by Demonstration: Learns tasks by observing a single demonstration and repeats them autonomously.

Personal Context personalizes Google apps (Google Forms, Google Docs, Gmail, etc.) based on user data.
Functionality:
- Merges data from various apps to provide personalized assistance.
- Operates privately under user instructions.
- Example: Gmail generating a smart reply using information from Drive, past emails, and Google Docs.

AI Mode enhances Google Search with AI-driven features.
Capabilities:
- Multi-tasking within search.
- Gathers links, images, and documents to present comprehensive search results.

AI model that generates music from text descriptions.
Functionality:
- Creates custom music based on text input (e.g., "Beethoven symphony").
- Generates royalty-free music for content creators.

SynthID secures content with invisible watermarks.
Function:
- Protects images and videos created by users.
- Includes a reader to verify if content is SynthID protected.
The invisible watermark is detectable and signals that the content is protected, preventing unauthorized use.

Satellite-Driven AI for Wildfire Detection:
- Detects fires in 270 square feet areas to prevent larger outbreaks.

$(270 square feet)$

AI-driven Drones for Disaster Relief:
- Delivers supplies to correct locations in disaster areas.
AI-Generated Driverless Cars:
- Autonomous vehicles for transportation.