If you use an AI chatbot and you’re concerned about your privacy and the personal information it ingests, you can do something about it. While Claude — our top pick for AI chatbots — is one of the better options when it comes to privacy, the default settings can still leave something to be desired if you value yours.
Luckily, you can mitigate or even eliminate the amount of data Claude knows about you, as long as you know where to look.
From having Claude train on your conversations to track your location and more, here are a series of settings you should consider changing if you want to stay on top of your privacy when using one of the best AI chatbots around.
More from CNET
Stop letting Anthropic train Claude on your chats and voice data
By default, most AI chatbots have a setting on by default that allows them to be trained on the conversations you have. This is the first and likely the most important setting you need to disable. Here’s how to turn it off.
The settings essentially remain the same across web, iOS and Android:
- Go to Settings.
- Go to Privacy.
- Toggle off Help improve our AI models.
- Toggle off Allow us to use your voice data.
You may not see this option because with some accounts, like Enterprise plans, Claude doesn’t automatically collect this data. So if you have a Claude account through work, you might not have to worry about this.
While you’re at it, if you’re a Claude Code user, these settings will let you delete Code sessions stored on Anthropic’s server without deleting local sessions on your computer.
If you’ve ever shared a chat with anyone, you can also delete those chats in this menu as well.
Read More: Keep Your Data Yours: How to Opt Out of AI Chatbot Training
Turn off location metadata collection
Anthropic will use your location when using Claude to “improve product experience.” It will use your device’s IP address to identify your city and region, but you can disable this setting — to some extent. How you go about doing this will depend on what device you’re using to access Claude.
From the web app:
- Click your Profile icon at the bottom left.
- Click Settings.
- Click Privacy.
- Toggle Location metadata to off.
According to the support article, if you’re using the mobile app (iOS or Android) or the web extension, the location settings are located within the device’s settings and not the Claude app settings. However, I found two ways to do this on Android. (See below.)
Mobile
The Claude app will not request access to your location unless your prompt has a specific location-based query, like “What are some of the best sushi restaurants around me?” Only then will the app request permission.
However, if you already have given Claude access on Android or iOS, here’s how you can revoke permission.
Android system settings
- Go to your phone’s Settings menu.
- Tap Apps.
- Tap Claude.
- Tap Permissions.
- Under Allowed, set the Location permission to Don’t allow.
Android Claude app
I was surprised to see a section in the Settings of the Android app that I didn’t see on iOS, but it will also allow you to disable your location right from the app.
- Open Claude’s Settings.
- Tap Permissions.
- Tap Location.
- Set it to Never.
iOS and iPadOS
- Open the Settings app.
- Scroll down to Apps.
- Tap Claude.
- Tap Location.
- Set Allow Location Access to Never.
One thing to keep in mind is that turning your location off can affect features like searches that require your location, but Anthropic explicitly states that it collects your IP address to infer coarse-grained location to ensure usage of Claude complies with its terms and to prevent abuse — despite questionable allegations.
Delete and disable memory
Like other AI chatbots, Claude has built-in “memory” so it can more easily have context about your conversations. Memory is on by default: The chatbot will remember personal preferences and other information and will be able to take that into consideration in other conversations, even if you don’t explicitly mention it in your current session.
If you’d prefer a clean slate, you can ask in the chat what memories it has collected about you and then tell Claude to delete specific memories or all of them. However, there’s a dedicated section in the settings to help you view and manage memories directly.
In the web version of Claude, do the following to access your memory settings:
- Click your profile icon.
- Click Settings.
- Click Your Claude in the Settings sidebar.
- To manage existing memories, click Your memory.
This will display all stored memories Claude has for you. You can edit a specific memory to your liking by clicking on the pencil icon on the right side of the memory or delete it entirely by clicking on the trash icon.
To prevent Claude from creating new memories from your conversations:
- Click on Your Claude in the settings menu.
- Toggle Generate memory from chats to off.
Use incognito mode for sensitive conversations
If you have had conversations with Claude that were sensitive in nature and still had the “Help improve our AI models” option enabled, odds are that Anthropic has already trained Claude on that conversation. There’s no way to change that.
However, you should find those conversations in your chat history and delete them. After you’ve done that and disabled AI training on your conversations, you should opt for incognito mode for any further sensitive conversations you might want to have going forward.
Incognito chats aren’t saved in your history, nor are they used to train models.