Google just announced its new Pixel 10 lineup at the Made by Google 2025 event in Brooklyn, NY .—other than the new Tensor G5 chip made using TSMC’s 3NM processes, there’s one thing that’s all over the new line of smartphones . And that’s AI . There’s Gemini in the photos app, there’s Gemini in the image editor, there’s AI voice translation, and it’s also sprinkled across the entire interface.
AI on smartphones so far has been about chatbots and prompts and actions. Google wants to bring a more proactive approach to things using the new system that it’s calling Magic Cue. Magic Cue uses the Tensor G5 chip and local LLM (Google’s Nano models) to automatically surface information in the right context.
Suggestions for texting and actions
Here’s Google’s pitch. Your friend texts to ask the address for tonight’s event. Pixel can pull the information from Gmail and bring up an option to share your reservation and address quickly. It can work for flight details too. This also works across many Google apps like Gmail, Calendar, Screenshots, Messages and more.
Credit: Google
As shown as a demo in the event, Google Pixel can help you do some other tasks as well. For example, if your partner asks you for a photo of your dog, it will show an option to bring up Google Photos showing just the photo of your dog.
Credit: Google
This is what’s meant by “contextual quiet computing.” It only shows up when it has some information to share, or an action to suggest—otherwise, it doesn’t bother you. Google says that you can control the data that Magic Cue has access to, and you can disable it at any time.
Magic Cue will also show up in Daily Hub, which you’ll find in your Discover feed. This new feature brings up a personalized digest showing what’s coming up in your calendar, relevant topics, playlists, emails, meetings, and more. All in one page.
Other AI features coming to Pixel 10
-
Live Translate: Gemini is also bringing live AI translation for calls, all of which happens on device. Pixel 10 translates the call in real time, and even matches the sound just like the speaker’s voice. So two people, talking in German and English, can carry on the conversation, in their native language, while Pixel translates the voice live. This works with English to and from Spanish, German, Japanese, French, Hindi, Italian, Portuguese, Swedish, Russian, and Indonesian.
-
Call Screen, but for missed calls: When you miss a call, the Take a Message feature will give you real-time transcripts. It will use AI to work out the next steps based on the message.
-
See more using Gemini Live: When you’re sharing your camera using Gemini Live, it can now highlight things in the camera view to make it easier to spotlight something that the AI really wants you to focus on.
