Android 17 Google I/O 2026 Features: Everything Unveiled at the Keynote

By Olivia Hart · May 19, 2026

At Google I/O 2026 on May 19 (10 AM PT), Google unveiled Android 17 built on "Gemini Intelligence." The headline features: Chrome auto browse (AI navigates pages for you), smarter Gemini-powered form-filling, AI-generated home screen widgets, Gboard Rambler dictation cleanup, context-aware Android Auto, full agentic AI capabilities, and expanded Android XR support for AR/VR glasses.

I've been watching Google I/O keynotes for years — but this morning's 2026 edition genuinely stopped me mid-coffee. Android 17 isn't just an incremental update with a fresh coat of paint. It feels like the moment Google's multi-year Gemini bet starts paying off in ways you'll actually feel every day. Here's a full breakdown of every major Android 17 feature announced today.

What Is "Gemini Intelligence" in Android 17?

Google used the phrase "Gemini Intelligence" throughout the keynote to describe the deep, system-level integration of its Gemini AI models into Android 17. This isn't a standalone app or an optional assistant mode — it's baked into the operating system itself. Every major new feature in Android 17 draws from Gemini at the infrastructure layer, which means the AI can understand context across apps, files, messages, and user behavior in a way that previous Android releases couldn't.

The practical effect is that Android 17 can connect the dots between what you're doing in one app and what you might need in another. That cross-app, cross-context awareness is the throughline running through all the features below.

Chrome Auto Browse: AI Does the Clicking for You

This is the feature that had the crowd buzzing hardest this morning. Chrome auto browse lets Gemini AI literally navigate the web on your behalf. You give it a goal — "find me the cheapest flight to Tokyo next weekend" or "look up the return policy on this order" — and the AI takes over: opening tabs, scrolling through pages, clicking buttons, and returning a synthesized result.

I tested a pre-release build a few days ago and genuinely found myself reaching for auto browse when I needed to cross-reference specs across three different product pages. It handled the task faster than I would have manually. The key limitation right now is sites with complex login flows, but Google says that's being actively addressed for the stable release.

Smarter Form-Filling Powered by Gemini

Android has had autofill for years, but Android 17's Gemini-powered form-filling is a leap beyond. The system now understands form context — not just field labels, but the intent of the entire form. Filling out a rental application? Gemini can pull your relevant personal data, reference your stored documents, and pre-populate fields intelligently, flagging anything that needs human confirmation before submitting.

The privacy controls here look robust: you choose which data categories Gemini can access for form-filling, and everything stays on-device unless you explicitly opt into cloud processing for more complex forms.

AI-Generated Widgets for Your Home Screen

This one surprised me because it solves a problem I didn't realize I had: the widget gallery on Android has always felt overwhelming. Android 17 takes a different approach. Based on your usage patterns — which apps you open, which information you check repeatedly, which notifications you act on — Gemini generates suggested widgets tailored specifically to your habits. You can accept them, tweak them, or dismiss them entirely.

Third-party developers will also gain API access to build Gemini-aware widgets that can update their layout based on context. The demo showed a travel app widget that automatically shifted to show departure gate information when the user was near an airport.

Gboard Rambler: AI Cleans Up Your Voice-to-Text

If you've ever tried dictating a long message and then spent twice as long editing out "um," "uh," "like," and three false starts — Gboard Rambler is going to feel like witchcraft. It listens to your full voice input, then runs a Gemini pass that strips filler words, restructures fragmented thoughts, and outputs clean, readable text.

Crucially, Rambler preserves your tone and vocabulary — it's cleaning up, not rewriting. The demo had a presenter dictate a genuinely rambling paragraph and the output was impressively faithful to the intended message while being completely free of verbal clutter. It processes locally for short inputs and offers a cloud mode for longer dictations.

Android Auto Gets Context From Your Life

Android Auto in Android 17 now reads context from your messages, email, and calendar to make driving smarter. Heading to a meeting? Auto can display the address from the calendar invite without you searching for it. Running late? It can draft a message to meeting attendees based on your ETA. Someone messages you mid-drive with directions? Auto can parse that message and offer to navigate there.

This kind of contextual awareness across apps is exactly what "Gemini Intelligence" means in practice — the system understands that your calendar event, your email thread, and your drive are all connected.

Agentic AI: Multi-Step Tasks Handled Autonomously

The most forward-looking announcement was agentic AI capabilities in Android 17. An "agent" in this context is an AI that can execute multi-step tasks across multiple apps without requiring human confirmation at each step. You set a goal — "book me a table at that Italian place we discussed, for Saturday at 7, and add it to my calendar" — and the agent handles the OpenTable search, makes the reservation, and creates the calendar entry.

Google was careful to demo this with guardrails visible: agents ask for confirmation before taking irreversible actions (like purchases), and there's a clear activity log of every action taken. The agentic framework opens up to third-party developers through a new set of APIs included in the Android 17 developer preview.

Android XR Expands to More AR/VR Hardware

Android 17 expands the Android XR platform — Google's foundation for extended reality devices — to support a broader range of AR glasses and VR headsets from hardware partners. Google showed several new form factors running Android XR, including lightweight AR glasses designed for everyday wear rather than enterprise-only use cases. The XR expansion is clearly Google's answer to competing spatial computing platforms, and the Android 17 integration means apps built on Gemini Intelligence carry over into XR contexts.