Android 17 Lands With Gemini Baked Into Everything

Google released Android 17 and Wear OS 7 on June 16, rolling out to Pixel devices first, and bundled them with a Pixel Drop where most of the new tricks run on generative AI. The headline change is inside the Gemini app: you can now edit video and create music tracks — from a text prompt or even an image — right inside a conversation. Under the hood that leans on Lyria 3 for music and the multimodal Gemini Omni, while the Pixel 10a picks up speech-to-translation built on AudioLM.

There's plenty that isn't AI, too. A new "bubble bar" parks your recent apps as bubbles at the bottom of the screen for faster multitasking. A screen-reaction mode records your selfie camera and your phone screen at the same time. Foldables get a gaming mode with a 50/50 layout and a dynamic on-screen gamepad. Wear OS 7 adds emergency detection — car crash, a fall, a missing pulse — that automatically pings your emergency contact, mirrors live app updates from the phone, and squeezes out up to 10% more battery. Google also wired Quick Share to talk to Apple's AirDrop on the Pixel 8a and 9a, and there's a new "Personal Intelligence" layer that connects your Google apps and chat history so Gemini answers with your context.

On the safety side, Find Hub gets a "Mark as Lost" mode, there's Live Threat Detection, and parents can set screen-time limits and content filters with a PIN — no Google account required for the kid's device.

Why a phone update is actually an AI shift

The interesting part isn't any single feature — it's where they live. When editing a video or generating a music track happens inside the stock Gemini app, not a downloaded creative tool, generative AI stops being a hobby and becomes the default way a phone works for billions of people. That's the same move Google pulled with search two decades ago: bury a powerful capability so deep into the everyday that nobody thinks of it as a separate thing. The Wear OS crash-and-fall detection is the quiet standout here — that's AI doing something that genuinely matters, not just making a meme.

For a regular person this cuts two ways. The upside: tasks that needed a separate app and a learning curve — trimming a clip, laying down a backing track — are now a sentence you type. The thing I'd watch: "Personal Intelligence" reaching across your apps and chat history is convenient precisely because it sees a lot of you. Worth a one-minute look at what you're actually handing it.

Google is doing what Google does best: shipping AI not as a destination you visit but as plumbing you stop noticing. A year from now the question won't be whether you use Gemini — it'll be whether you remember which of your phone's everyday actions quietly became a model call.