How to enable and use the "Memory" feature in Google Gemini bot
Users have been heavily engaged with bots since the AI boom in late 2022, especially with the revolutionary steps taken by major tech companies. With Google launching the “Gemini” bot , which directly competes with “ ChatGPT ,” some similar features have emerged between them; perhaps the most recent of which is the “Memory” feature that “OpenAI” introduced last July with “ChatGPT,” and is now also available in “Gemini.” For those who don’t know, this feature enables the bot to remember personal information that users share during conversations and use it to provide personalized responses relevant to their needs. Below, we share with you how this feature works in “Gemini,” how it can be used, and whether it raises any privacy concerns.
Google explained in an update notice on the Gemini website that the bot can now remember your interests, personal preferences, or even your aspirations using the new “memory” feature, which results in gradual improvements over time to help provide more consistent and accurate responses to your needs. This feature, which OpenAI previously introduced with ChatGPT, works in two ways: The first is automatic. For example, if you previously told the bot that you prefer to receive structured responses in the form of brief bullet points, Gemini will take that into account in all future responses. Or if your little one loves a type of animal, the bot will remember these details when you ask it to prepare a message for your child for his birthday, for example.
The second way is to ask the bot to remember specific details by typing simple phrases like “Remember [this]” or “Don’t forget that I am [this].” There’s plenty of room to add things you want Gemini to remember, whether it’s about your work, hobbies, or general life aspirations. Ultimately, the goal is to make the bot more capable of delivering personalized responses that suit you and make it easier to interact with without having to re-explain the same details every time. It’s like giving Gemini a user manual, designed by you.
Of course, the user has full control over what the bot remembers. Any information you have shared with Gemini can be easily edited or deleted, and you can also ask it to forget specific details from conversations. Unfortunately, you will need to subscribe to the paid version of Gemini, known as “ Gemini Advanced ,” to use the “Memory” feature, keeping in mind that it currently only supports English.
If you are subscribed to the paid service for the Google Gemini chatbot, you can take advantage of the memory feature and try it now through the robot's website , as Google indicated that the feature is still not currently available in the "Gemini" application for Android and iPhone. So after heading to the site, click on the gear icon (⚙️) in the sidebar, then click on "Saved info" from the menu, or alternatively go directly to the "Saved info" page and activate the activation button at the top. From this page, you can include any information you want to share with the robot, by clicking on the "Add" button, or manage the memories that the robot records, such as deleting information by clicking on the menu button ( ⁝ ) in front of the saved text and choosing "Delete" or "Edit" if you want to edit the text.
You can also ask the bot directly to save specific information such as food preferences, preferred languages, and work information during the conversation. All you have to do is start the prompt by typing words like “Remember,” “Keep in mind,” “I prefer,” etc. Your phrases can vary depending on the context, but in general, it’s best to use simple language and avoid too much jargon. At the end, Gemini will tell you that the information has been saved.
Now you can use the bot as you normally would and see the results for yourself. If you ask a question and the answer requires the information it has on you, the bot will let you know by adding a “Your saved info” notification in the “Sources and related content” section below the answer.
True, this feature raises privacy concerns, but Google has confirmed that it does not use users’ conversations or their saved information to train the model, so there is no need to worry. But if you are still concerned about this aspect, you can disable the feature completely.