Google’s AI glow-Up: Here are all the game-changing announcements from I/O 2025

Google just dropped a whole lot of AI at this year’s developer conference, I/O 2025. From smarter Search and a next-gen Gemini to AI-generated videos and new smart glasses, Google is going all in on artificial intelligence — and how it’s going to power your day-to-day tech in the future.

Here’s a full rundown of everything major that was announced, explained:

Gemini Live is going global, and now it sees what you see

Gemini Live, Google’s visual AI assistant, is now rolling out for free on both Android and iPhones. It’s basically a real-time AI that can use your camera to understand what’s in front of you — whether it’s your homework, a product, or some random object — and then respond like a human.

It’s powered by the same tech from last year’s Project Astra and lets you ask follow-up questions, making Search feel more natural and interactive.

AI Mode hits Search with chat-style smarts

Google’s rolling out a new AI Mode tab inside Search. It lets you talk to Google like you would with a chatbot. Ask follow-ups, dig deeper, and actually talk to your search engine.

AI Mode can also make custom charts for things like finance or sports stats, and will soon personalise results using your Gmail and previous searches. It’s Google Search but way smarter. It will be initially available to users in the US only starting this week, but will roll out globally in phases.

Gemini joins Chrome to help you browse better

Gemini will now be baked directly into the Chrome browser. Gemini Pro and Ultra subscribers can use a new Gemini button to summarise content, break down complicated pages, or help you navigate websites with less stress.

Later this year, it’ll even be able to handle multiple tabs at once. Think of it like your personal web butler.

Say hello to Flow — your new AI Studio

Flow is a new AI filmmaking app that can make short video clips just from text prompts or pictures. It pulls power from Google’s Imagen, Veo, and Gemini models, letting you create eight-second videos and stitch them into longer stories.

It’s perfect for creators, marketers, or anyone who wants to make cool video content without having to edit for hours.

Imagen 4 and Veo 3: AI photo and video tools level up

Imagen 4 is Google’s latest text-to-image tool and it’s way better at handling text inside images. It also supports more formats like square and landscape. Veo 3, the latest version of Google’s AI video model, now adds realistic audio too. So it’s not just what you see — it’s what you hear.

Veo can generate ambient sounds, voice lines, and even sync lip movements to match your characters. It’s available now for Gemini Ultra users via the Flow app and Vertex AI.

Gemini 2.5 Pro goes deep with ‘Deep Think’

Google’s most powerful AI model just got an experimental feature called Deep Think. It’s made for tougher stuff like maths and coding and can evaluate multiple solutions before giving you the best one. It’s still in early testing, but shows how far reasoning models are going.

The lightweight Gemini 2.5 Flash model also just launched for everyone through the Gemini app and Google AI Studio.

Android XR is gaining serious momentum

Google and Xreal are teaming up for Project Aura, a new smart glasses prototype using Android XR. The glasses come with Gemini built-in, a huge field of view, and what looks like built-in mics and cameras.

Google’s also working with Samsung, Warby Parker, and Gentle Monster to build more Android XR glasses. So yeah, the wearable AI race is heating up.

Gmail gets way better at writing your replies

Google is upgrading Gmail’s smart reply tool. Soon, it’ll scan your inbox and Google Drive to suggest responses that sound more like you. It’ll also adjust the tone depending on who you’re talking to — casual for friends, formal for your boss.

This upgrade is launching in July for web, Android, and iOS via Google Labs.

Shopping gets smarter with AI

Google’s AI shopping tools are getting a glow-up. You’ll be able to upload a full-body pic to see how outfits look on you. Google’s AI understands how clothes fall and move, so it gives you a proper preview.

There’s also an upcoming feature called “agentic checkout” that can literally buy things for you. Welcome to the AI shopping assistant era.

Chrome will help you fix weak or hacked passwords

Later this year, Chrome will start helping users automatically replace compromised passwords. If your password is flagged as unsafe, Chrome will suggest a stronger one and even update it on your behalf — but only after asking for your permission.

It’s a solid step toward better security without the headache of managing dozens of logins yourself.

Ultra AI is here — but it’s not cheap

Google’s launching a $250/month subscription called Google AI Ultra. It gives users access to the most powerful Gemini models, high usage limits, and early access to experimental tools like Project Mariner, which can complete up to 10 tasks at once.

It’s clearly aimed at enterprise users, developers, and serious power users who need that extra AI firepower.

That’s the biggest stuff from I/O 2025. Whether you’re building content, browsing the web, or just trying to make your life easier, Google is betting big on AI to do it all. The tech might still be evolving, but this year’s updates show Google is dead serious about leading the AI race.