Google I/O 2025 highlights: Gemini, AI Search, and XR take centre stage

From major Gemini upgrades to new AI-powered Search features and 3D video calling, Google I/O keynote event detailed plans for Google's AI-first future

Google kicked off its annual developer conference on May 20 with a keynote event where artificial intelligence was front and centre. The company unveiled major upgrades to its Gemini AI models, previewed new agentic AI capabilities, introduced tools for AI-enhanced Search, and gave a glimpse into its evolving Android XR platform for headsets and smart glasses.

 

Here are the top announcements and takeaways from the Google I/O 2025 keynote:

 

Updates to Gemini app:

 

Google announced that Gemini Live’s camera and screen-sharing capabilities, based on Project Astra, will begin rolling out to Android and iOS users this week. The feature allows users to have real-time conversations with Gemini about what's visible on their screen or through their phone’s camera.

 

Gemini Live is also being integrated more deeply into Google's app ecosystem. Soon, users will be able to fetch directions from Maps, set reminders, and interact with other services using Gemini’s conversational interface.

In addition, Gemini’s Deep Research mode is getting an upgrade, enabling users to upload files and images to receive more contextual and detailed responses.

 

Veo 3 and Imagen 4 models:

 

The keynote also saw the debut of two new generative models: Imagen 4 and Veo 3. Imagen 4 is Google’s next-generation text-to-image model, which now produces visuals with significantly greater detail and improved rendering of text within images. Veo 3, Google’s latest text-to-video model, takes video generation further by rendering fluid motion graphics and generating native soundtracks that match the visuals.

 

AI Mode for Search:

 

Google also detailed new enhancements to AI Mode, its evolving AI-powered Search experience. AI Mode will now offer more interactive, personalised, and visually rich results. For example, users will be able to search for complex topics such as finance or sports statistics and receive visualised results in the form of AI-generated graphs. In the case of shopping queries, AI Mode will power features like “try it on” for apparel, helping users preview how clothing might look.

It will also introduce more personalisation by offering results influenced by past searches and enabling connections to other Google services—starting with Gmail—for deeper contextual content. The company confirmed that AI Mode is rolling out to all users in the US this week, with plans for expansion later.

 

Android XR:

 

Google also took the wraps off its Android XR platform, showcasing how it plans to bring mixed reality experiences to smart glasses and headsets. In a live demonstration, the company showed how Android XR powers smart glasses for everyday use cases such as messaging, navigation, and real-time language translation using built-in lens displays.

 

Google has partnered with eyewear brands including Gentle Monster and Warby Parker to develop Android XR-based smart glasses, with more announcements expected in the near future.

 

Agentic AI experience:

 

Another major theme at the event was agentic AI—systems that can carry out multi-step tasks on a user’s behalf. Through its Project Mariner experimental agent, Google demonstrated how users could complete tasks like booking tickets or making online purchases entirely through AI, without needing to manually visit any websites. Similarly, the Gemini app is getting an Agent Mode, which will allow it to retrieve and summarise relevant web data in real time, simplifying workflows that would typically require multiple apps and steps.

 

Beam 3D video communication:

 

At the keynote event Google also previewed its new 3D teleconferencing technology called Beam. It uses a combination of hardware and software—including a six-camera array and a custom light field display—to create a highly realistic 3D representation of the person on the other end of a call. The result is an immersive video experience that mimics the feeling of an in-person meeting. Google confirmed it is working with partners like Zoom and HP to bring Beam to enterprises, with HP expected to launch the first Beam-enabled devices soon.