Google Gemini Removes ‘Ask About Screen,’ Gains Automatic Screenshot Analysis – But There’s A Catch

Google is testing a new Gemini feature that lets the AI assistant inspect what appears on a device screen and answer related queries automatically. The tool, now in limited testing on select phones, aims to make help sessions faster by removing the need for users to share screenshots manually.

The upgrade builds on Gemini’s current support for files such as images, videos and documents, and extends that awareness to live on-screen content. Early testers report that it already works in some everyday tasks, though accuracy varies and Google is expected to adjust the system using test feedback.

How Automatic Screen Analysis Works

Inside Google, the capability is known as screen context and is designed to detect when a question relates to what is currently open. Instead of asking people to press an “Ask about screen” shortcut, Gemini quietly takes a screenshot in the background and then uses that capture while drafting its response.

Once active, the Gemini feature can reference whatever appears on display while answering queries. It can break down details in an image, help explain steps in a video, or give tips within an app interface itself. The behaviour is somewhat comparable to Gmail’s reminder when someone forgets to attach a file that is mentioned in an email.

Rollout, Testing and Availability

The Gemini feature first showed up in development versions earlier this year, but it is now surfacing on a limited number of consumer devices. Those who receive early access see an onboarding message that describes the tool, highlights the use of screenshots and allows people to switch the feature off immediately if preferred.

Google has not shared a schedule for wider public availability and has not confirmed which regions will see the next phase. The company is expected to introduce the feature over the coming months in stages rather than a single global release, following typical phased rollout patterns for major Assistant changes.

Privacy Controls and User Feedback

Privacy controls are built into the Gemini feature, and Google says users keep control at all times because the option can be disabled whenever required. When turned off, Gemini stops capturing new screenshots for context, which may reassure people who prefer that on-screen information stays outside the assistant’s view.

Some of the first people with access have posted reactions on Reddit and X, sharing how the Gemini feature behaves with different prompts. They reported that the assistant sometimes correctly spots when a question refers to content on display, yet in other attempts it fails to link the two, especially when the wording of the query is less direct or clear.

Feedback gathered during this early phase is expected to guide refinements before wider release. Testers suggest that improved detection of vague queries could be helpful, along with clearer prompts that show when Gemini has used a fresh screenshot. Google has not yet detailed specific adjustment plans for the Gemini feature.

One message seen by testers also includes the line: “Survey ✅ Thank you for completing the survey!” which appears to relate to Google’s internal research around the Gemini feature. The presence of such notices underlines that Google is still treating this screen context ability as experimental.