At Google I/O 2025, the technology behemoth introduced a range of innovative features for Gemini Live, its AI-driven assistant stemming from Project Astra. These new capabilities signify a monumental advancement in real-time, multimodal AI communications, drawing users nearer to a genuinely intelligent digital assistant that can perceive, hear, and act on their behalf.
Here’s a summary of the six most thrilling new features making their way to Gemini Live:
1. Complimentary Camera and Screen Sharing for Everyone
One of the most revolutionary updates is the introduction of free camera and screen-sharing functions for all users on Android and iOS. This allows anyone to point their device at an object or share their screen, enabling Gemini to interpret visual data in real time.
Whether you’re fixing a gadget, identifying greenery, or reviewing a document, Gemini can now “see” what you see and provide intelligent support. This feature, which was previously exclusive to premium subscribers, is globally rolling out starting May 21, 2025.
2. Effortless Integration with Google Applications
Gemini Live is becoming more integrated with Google’s suite of apps. It will soon interface with tools like Google Maps, Calendar, Tasks, and Keep. This integration allows users to request Gemini to schedule events, set reminders, or navigate—all within one conversation.
For instance, you could say, “Show me my agenda for tomorrow and add a reminder to call Mom at 3 PM,” and Gemini will manage it across various applications.
3. Live Web Browsing and Document Interaction
In a live demonstration, Google showcased Gemini Live assisting a user in repairing a bicycle. The assistant searched online for a user manual, opened it, and navigated to the relevant brakes section—all in response to natural voice queries.
This highlights Gemini’s capability not just to locate online resources, but also to interact with them in real time. It can browse the internet, open documents, and extract specific details, making it an invaluable tool for research and productivity.
4. Advanced Multimodal Context Awareness
Gemini Live has now incorporated sophisticated multimodal interactions. In one instance, a user displayed a tray of bolts using their phone camera and asked Gemini to cross-reference the items with an email in Gmail. Gemini accurately identified the correct hex nut size and spotlighted it in the live video feed.
This level of context-aware intelligence—melding visual data with personal information—sets a new benchmark for AI assistants.
5. Background Task Management and Phone Calls
Another remarkable feature is Gemini Live’s capability to make phone calls on your behalf. In the demonstration, a user instructed Gemini to call a nearby bike shop to inquire about a part. The AI executed the call in the background, allowing the user to continue engaging with Gemini for other tasks.
After the call concluded, Gemini returned with the shop’s reply and even offered to place a pickup order. This type of agentic behavior—where the AI operates independently to fulfill tasks—marks a significant move towards fully autonomous digital assistants.
6. Multi-Speaker Recognition and Conversational Continuity
Gemini Live can now manage interruptions and multiple speakers without losing track of the discussion. During a demo, while Gemini was on a call, another individual posed a question to the user. Gemini paused, awaited the user’s reply, and then smoothly resumed the initial conversation.
This situational awareness guarantees that Gemini can operate naturally in real-world settings, such as bustling homes or offices.
Bonus: Context-Sensitive Shopping
In a final showcase, the user queried Gemini to assist in finding a dog basket for their bike. The assistant utilized contextual hints—like recognizing the user’s dog from Google Photos—to recommend appropriately sized baskets. Although the demo didn’t demonstrate a complete transaction, Gemini successfully offered to arrange a pickup order for a bike part earlier, hinting at future e-commerce functionalities through Project Mariner.
When Will These Features Be Released?
Google is presently testing these features with trusted users and intends to gradually deploy them across Android smartphones and XR (extended reality) devices. While no precise timeline has been provided, the company underscored its dedication to privacy and user input during the rollout.
Why It Is Significant
These enhancements position Gemini Live as a credible contender in the AI assistant market, competing with offerings from OpenAI, Apple, and Microsoft. With the capability to perceive, hear, and act in real-time, Gemini Live is transforming from a passive assistant into an active partner.
As AI progressively integrates into our everyday lives, tools like Gemini Live are paving the way for a future where digital assistants are not merely beneficial—they’re indispensable.
Stay tuned as Google keeps refining and broadening Gemini Live’s functionalities, bringing us nearer to the era of genuinely intelligent, multimodal AI.
