Google’s Pixel 9 line is powered by cutting-edge hardware like the Tensor G4 processor and plenty of RAM, designed to keep your phone feeling fast and fresh for years to come. But all of that hardware is also designed to power brand new AI experiences.

“Android is reinventing your phone with Gemini,” wrote Sameer Samat, Google’s president of the Android ecosystem, in a blog post published Tuesday. “With Gemini deeply integrated into Android, we’re rebuilding the operating system with AI at its core. And redefining what phones can do.”

Here are the big new AI features coming with the new Pixel devices.

Gemini Overlays and Gemini Live

Gemini, Google’s AI-powered chatbot, will be the default assistant on the new Pixel 9, Pixel 9 Pro, Pixel 9 Pro XL, and Pixel 9 Pro Fold phones. To access it, just press and hold your phone’s power button and start speaking or typing your question.

A big new addition is that you can now bring up Gemini on top of any app you’re using to ask questions about what’s on your screen, for example to find specific information about a YouTube video you’re watching. You can also generate images directly from this overlay and drag and drop them into the app underneath, as well as upload a photo to the overlay and ask Gemini questions about it.

If you buy the more expensive Pixel 9 Pro (starting at $999), Google bundles a free year of the Google One AI Premium plan, which normally costs $19.99 per month and gives access to 2TB of cloud storage and Gemini Advanced, which lets you try out Gemini directly in Google products like Gmail and Docs to summarize texts and conversations.

Crucially, Gemini Advanced also offers access to Gemini Live, which Google describes as a new “conversational experience” that makes speaking to Gemini more intuitive (don’t worry, I’m not the only one who struggles to keep track of all the things Google calls “Gemini”). You can use Gemini Live to have natural conversations with Gemini about whatever’s on your mind. Google says you can also use it to help with complex questions and interviews, choosing between a variety of voices that sound amazingly lifelike, according to demos Google showed Engadget earlier this month.

Recently, OpenAI released Advanced Voice Mode, a similar feature for paying ChatGPT customers with a voice assistant that can speak, sing, laugh and reportedly understand emotions. When asked if one of Google’s goals was to make Gemini Live sound as human as possible, Sissie Hsiao, vice president and general manager of Gemini Experiences, told Engadget that Google “is not here to show off the technology. We’re here to build a super helpful assistant.”

Photos and camera functions

Google is using AI to significantly improve the way you take and edit pictures with the Pixel 9 phones, something the company has been focusing on for years. For example, a new feature called “Add Me,” released as a preview with the new devices, lets you take a group photo and then separately take a picture of the photographer and seamlessly add it to the main image – handy if you don’t have anyone to take a picture of your entire group.

Meanwhile, Magic Editor, Android’s built-in AI-powered editing tool, can now suggest the best crops and even enhance existing images by adding details using generative AI to reveal more of the scene. And finally, the new “Reimagine” feature lets you add elements like autumn leaves or making the grass greener – while it can spice up your images, it blurs the line between real and fake memories.

Circle to Search now enables sharing

You can already search for anything you see on your phone by simply circling it, but now AI intelligently cuts out anything you’ve circled and allows you to instantly share it in a text message or email. Convenient.

Pixel screenshots

If you don’t know how to sort through the reams of images of receipts, tickets and social media screenshots that are flooding your phone’s photo gallery, you can call on AI to help. A brand new app called Pixel Screenshots, available on the new Pixel devices at launch, scans your photo library (after you give it permission), selects screenshots and then identifies what’s in each image. You can also click pictures of real-world signs (for example, from a music festival you’re planning to attend) and ask the app relevant questions directly, like when tickets for the festival go on sale.

Call notes

A new feature called Call Notes automatically saves a private summary of every phone call. This lets you refer back to a transcript later to quickly look up important information from the call, such as a meeting time, address or phone number. Google notes that the feature runs entirely on the device, meaning nothing is sent to Google’s servers for processing. And everyone on the call will be notified if you have Call Notes enabled.

PixelStudio

We’ve been able to use AI to generate images for a long time, but Google is finally integrating the feature directly into Android thanks to Pixel Studio, a new dedicated image generation app for Pixel 9 devices. The app runs on both a device model with the new Tensor G4 processor and Google’s Imagen 3 model in the cloud. You can share any images created in the app directly via message or email.

A similar feature called Apple Image Playground is coming to newer iPhones with iOS 18 in September.

Custom weather reports

Google will use artificial intelligence to create custom weather reports for your specific location right at the top of a new Weather app, so you “don’t have to scroll through a bunch of numbers to get a sense of the day’s weather,” the company said in a blog post.