Google is rolling out a new Memory summary in its Gemini AI assistant that shows users exactly what the chatbot has learned about them, and lets them correct or delete anything it has got wrong. The feature began reaching users on 9 November, and it turns Gemini's scattered memories into a single readable page.

Until now, Gemini has quietly remembered details from conversations and used them to tailor its answers, but users had no simple way to see the full picture. The new summary collects what the assistant understands about a person and displays it in grouped sections. The categories include The Essentials, described as a snapshot of core background, location and key relationships, along with Career and Professional Projects, Travel and Daily Gear, Media and Entertainment, and Fitness and Lifestyle. Each section contains bullet points with specific details drawn from earlier chats.

Users can reach the page from Settings, then Personal Intelligence, then Memory summary, or by opening the personalisation settings page for Gemini on the web. They can also simply ask Gemini what it knows about them. A text field at the bottom of the page lets people add or remove details in plain language, for example by asking it to remember that they speak Japanese, or to remove an old phone number from its memory. Google says Gemini may ask follow-up questions if a request is ambiguous, so the summary stays accurate.

Edit, delete and check the history

The controls go further than a single list. A pencil icon opens broader editing, and a menu option labelled Memory history shows when specific pieces of information were added or changed. That lets a user fix an out-of-date job title, a wrong address or a preference they no longer hold directly. The feature is built on Gemini's existing memory capabilities, but it is the first time Google has presented the stored details back to users in this organised, editable form.

The rollout comes with limits. Officially, Memory summary is available to Google AI subscribers in the United States, using English, who are aged 18 or older and have both Keep Activity and Memory switched on in their account settings. Google has not said when it will expand to other languages or to more countries. The feature is appearing first on the web, and 9to5Google reported that it had not yet seen it in the Gemini app on Android or iOS, although the experience there should be similar when it arrives.

Wider access than the official limits suggest

Early testing suggests the real-world rollout may already be wider than Google's own description. Android Authority reported finding Memory summary working on Gemini's website with a free account outside the United States, including through a mobile browser, even though it was not visible in the Gemini mobile app. Gadgets360 similarly reported being able to use the feature on a free account. That is a familiar pattern for Google, which often switches features on for slices of users before widening access over days or weeks, so availability may vary from person to person for now.

The launch reflects a broader shift in how AI assistants handle personal data: as chatbots remember more, the ability to inspect, correct and delete those memories becomes a privacy question as much as a convenience. A summary that is wrong about a person's job, location or interests could shape every answer that follows. The practical advice is simple: open the Memory summary, read what Gemini thinks it knows, and remove anything wrong or that you would rather it forgot.