Hovimo
Technology

Google Gemini allows viewing and editing of personal AI memory

Google has introduced a new feature that allows users to manage data stored in the memory of the Gemini chatbot for personalized responses.

⏱ Story timeline · Put Gemini to work for you

Editing the AI’s memory is one way you can shape its responses.
Editing the AI’s memory is one way you can shape its responses. — Engadget
Why it mattersThe ability to control the data that the AI knows about the user increases privacy and trust in the technology.
Executive summary across all sources
  • Users can view and edit Gemini's memory via Personal Intelligence settings.
  • For Gemini's memory to work, the Keep Activity setting must be enabled.
  • Google has introduced the Gemini 4 Neon model for rapid data analysis.

How Gemini Memory Works

Google has updated the Personal Intelligence feature in Gemini, allowing users to see and edit information that the AI accumulates about them during conversations 1. This memory helps the bot better adapt its responses by taking into account the user's preferences and personal data, such as career or interests 1. For this feature to work, the "Keep Activity" setting must be enabled 1.

A finger tapping on a Gemini app icon.
A finger tapping on a Gemini app icon. — Engadget

The implementation of Gemini's memory system is currently restricted to users in the United States who subscribe to a paid Google AI plan 1. To manage these settings, users navigate to the gear icon in the Gemini web app, select Personal Intelligence, and access the Memory summary 1. While the system tracks a history of all manual edits made to the profile, users can also issue direct prompts to the AI to add or subtract specific details 1.

Gemini Memory Management
Key user actions for controlling personalization data
✓
Turn on the Keep Activity setting
Completed
✓
Перейти до налаштувань Personal Intelligence
Completed
✓
View and edit AI memory
Completed
From the article · Google

Furthermore, the recent introduction of the Gemini 4 Neon model enhances the AI's ability to extract data from receipts or tables and integrate that information more effectively with a user's Gmail or Google Docs 2.

Gemini Data Management
Changes to saved information controls
Auto-save
The AI accumulated information without the ability to easily view and manually edit it.
Transparent control
Users can view, delete, or modify facts, as well as disable memory.
From the article · Google

Data Control

Users can review the Gemini memory report through the settings menu, where information is grouped by categories 1. The editing feature allows users to delete or modify individual facts, similar to editing a document 1. Additionally, users can completely disable the memory system or delete chat history, which limits the AI's ability to use this data for personalization 1. At the same time, Google introduced the Gemini 4 Neon model, which provides faster document analysis and better integration with Google services such as Gmail 2.

The Memory summary windows in Gemini's Settings menu.
The Memory summary windows in Gemini's Settings menu. — Engadget
Background

Artificial intelligence is becoming increasingly personalized, which requires new tools to protect user privacy.

What’s next

Further expansion of privacy control features in the Google AI ecosystem.

Story chronology & developmentFull timeline →
  1. Google DeepMind Updates Project Genie with Genie 4 and Chainable Worlds

    Google DeepMind has released Genie 4, a new interactive 3D world model featuring real-time text editing, native audio, and community remixing tools.

  2. Google Gemini allows viewing and editing of personal AI memory

    Google has introduced a new feature that allows users to manage data stored in the memory of the Gemini chatbot for personalized responses.

  3. Gemini 4 Neon Takes Top Coding Rank Amid Google's Rapid-Fire Model Sprint and Pricing Paradox

    Google introduced Gemini 4 Neon on November 18, 2026, positioning it between Hydrogen and Carbon for complex coding, enterprise analysis, and science workflows. While the model claims #1 on the Artificial Analysis Coding Agent Index and strong cybersecurity scores, its launch exposes temporary pricing inversions, a dropped sibling model, and heavy reliance on company-reported evaluations.

Written with AI from the cited sources and checked against them. Open the sources to verify.

Original report ↗