Google is expected to unveil several new Gemini AI features at the Google I/O event in mid-May 2026, but the company already shared its AI innovations a week earlier, hosting the I/O 2026 edition of “The Android Show” on YouTube to announce new Android 17 features and the Android-based Googlebook laptops that will replace Chromebooks. The AluminumOS laptop makeover was already expected, as was the focus on AI.
Indeed, Google said on stage that Google Books were designed for Gemini Intelligence “from the ground up,” sharing a few key features that should set these computers apart from Mac and Windows devices, including Magic Pointer and support for creating custom widgets. Magic Pointer is easily one of the best features to come out of Google’s pre-I/O event, a reinvention of the cursor for the AI era that can accelerate AI workflows. Unfortunately, AI’s reinvention of the classic mouse cursor is the latest development in this somewhat boring trend: putting AI everywhere on mobile and PC and hoping some of those features work, while ignoring user privacy concerns.
Magic Pointer, based on AI Pointer technology from Google’s DeepMind, allows users to interact with AI directly in their workflows, without having to save content and move it to the Gemini app or write a prompt. To summon the AI to where the mouse cursor is, you shake it, at which point Gemini will understand your context, whether you’re looking at an image or text, and offer suggestions. The user can still use prompts, and DeepMind demos show that the AI Pointer works with voice inputs. Perhaps the closest thing to Magic Pointer is Google’s Circle to Search feature on Android smartphones, which allows users to perform various AI-based searches by simply drawing a circle on the screen.
Why Magic Pointer is awesome
Google’s announcement about Googlebooks lacked expected details for a new product category expected to hit stores this year. Google has not mentioned any specifications for Googlebook devices from different PC manufacturers, nor has it revealed prices. These are important details considering the competition. The $599 MacBook Neo has been a huge success for Apple this year. At the same time, Microsoft was forced to increase prices on some of its Surface devices. As a reminder, shortages of memory and storage chips, linked to the demand for AI, are hurting smartphone and laptop suppliers.
In this context, Magic Pointer can be an awesome tool to speed up work on a laptop without requiring hardware resources, like a faster chip and more RAM. Indeed, a movement of the mouse brings up Gemini, which sees what the user sees and understands the context around the cursor. AI can recognize that you are working with a word processor, looking at images (and what they show), or coding. After that, the AI can offer suggestions and follow your cursor movements to perform actions.
Google’s demonstrations included pointing Magic Pointer at an email, prompting Gemini to suggest actions for Google Calendar, Gmail, and Google Maps. Using the magic pointer to select multiple images on the screen informed Gemini that the user might like to view all three images together, and when the user clicked on this suggestion, the AI created a text prompt for the task and began generating the image. “To do this on today’s laptops, you have to right-click, save images, upload them to a chatbot, then ask,” Google said. “Instead, your Googlebook can understand the context and get to work right away.”
Why Magic Pointer is concerning
The Google DeepMind team showed how easy it is to use voice to interact with Gemini while using the AI Pointer functionality, without being specific like in text prompts. Vague interactions such as “Fix this” or “What does this mean?” » would work in human-to-human communication, because people use gestures to point to their surroundings. The technology developed by DeepMind for AI Pointer and Magic Pointer works similarly. In one example, the user points to an image containing the name of a train station and a photo showing a place of interest, telling the AI to navigate between the two locations. Gemini recognizes both addresses and opens a Google Maps window that displays directions between the two locations. This experience is much faster than performing the steps manually.
But people who worry about their privacy on devices that put AI at the forefront would be right to question the privacy implications of Magic Pointer. Google hasn’t yet explained what the AI can see when you shake the slider, or when data collection starts and stops. Additionally, Google does not specify whether Magic Pointer data processing takes place on the device or whether the data is sent to the cloud. Finally, Google does not explain whether any of this data, which can be sensitive and very personal, is ever used to train future Gemini models or even to improve Magic Pointer.
These are things to keep in mind before using new AI features, no matter how cool they may be. That said, you won’t need to wait for Googlebooks to test Magic Pointer, as a limited version of the feature works in Gemini in Chrome on Mac and Windows PCs. Users can also test AI Pointer in Google AI Studio.
