Chatbots are a dime a dozen and people continue to rely on them for more than just casual conversations at this point. Whether you’re vibe coding an app, generating images or video, or something in between, there’s a reason people are leaning on the likes of Gemini, Claude and ChatGPT for more than what you’d get from a search engine.
All that said, AI itself is even hungrier for your data than the tech giants that create it. And Google, OpenAI and others use your data to help make their AI models better — unless you do something about it. By default, most chatbots will use your data to train models, so there are some settings you should look into if you want to keep your conversations as private as possible.
Below, we’ll show you how to disable AI model training on some of the most popular chatbots and delete the conversations if you want to start with a clean slate.
Gemini
You’ll need to turn off a few settings in your Google account and in Gemini to prevent your data from being used to train AI models.
Turning off Web & App Activity will prevent your new searches, location and other data from being recorded to your Google account and used for AI training. However, turning this off does not delete the existing activity already on your account. You’ll still need to click the delete button on the same page and select to delete your data from either all time or your specified range.
Gemini
Gemini has its own settings that you’ll need to disable in order to prevent Google from using your data for model training:
- Log in to Gemini.
- Click the Settings cog.
- Click Activity.
- Click the drop-down and set it to Off.
Similarly, you’ll also need to manually delete the existing Gemini activity (your conversations) by clicking the delete button and selecting a time range.
Claude
If you’re looking to remove your data or prevent Anthropic from using your data in Claude, it’s a little more straightforward.
Change Model Improvement settings:
- Log in to your Claude account.
- Click your profile icon.
- Click Settings.
- Click Privacy.
- Set the toggle for Help improve our AI models to off.
While you’re here, you might also want to toggle the Location metadata off for good measure.
Deleting your conversations in Claude is also easy:
- Log in to your Claude account.
- Click Chats on the left sidebar.
- Click Select Chats.
- Click Select all.
- Click Delete.
Per Anthropic’s privacy policy, deleted chats are stored for up to 30 days in its back-end systems. If you’ve previously allowed Anthropic to use your data, and you turn the setting off, new and existing chats won’t be used to train models going forward, but your data will be used in training that is in progress and won’t be removed from models that have already been trained.
ChatGPT
ChatGPT is the world’s most-used chatbot. If you’ve been using it for a while without adjusting your settings, you’ve probably given OpenAI a lot of data.
Turn off model improvement in ChatGPT:
- Log in to your ChatGPT account.
- Click your profile icon.
- Click Settings.
- Click Data controls.
- Click improve the model for everyone.
- Switch the toggle to improve the model for everyone off.
This pop-up menu contains two additional settings to include your audio and video recordings, but toggling the main switch at the top will disable both of these settings.
To delete all of your chats, there’s a dedicated button in the Data controls menu. Like Anthropic’s policy, OpenAI also stores deleted chats for up to 30 days.
(Disclosure: Ziff Davis, CNET’s parent company, in 2025 filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)
Meta
The TL;DR for trying to get Meta to stop using your data to train its AI is that there’s no guarantee. There’s no elegant solution, and you’ll have to submit a form, which Meta may take action on.
Making your profiles private and limiting your posts will shield you the most from AI training. Or you can just delete your Instagram and Facebook profiles.
Instagram users looking to get away from the fact that others can reuse their own media can opt out to avoid the deepfake apocalypse.
Aside from that, all else you can do is mute Meta AI.
I was unable to find the specific setting to get Meta to stop using your data, and Meta did not respond to a request for comment when asked for a guaranteed way to opt-out of AI model training.
Grok
It’s almost surprising that Grok will allow you to turn off AI model training on X, but you can. Here’s how.
- Log in to your X account.
- Click the More button.
- Click Settings and privacy.
- Click Privacy and safety.
- Click Grok & Third-Party Collaborators.
Uncheck the box next to “Allow your public data as well as your interactions, inputs, and results with Grok and XAI to be used for training and fine-tuning.”
While you’re at it, you might want to uncheck Allow X to personalize your experience with Grok and Allow Grok to remember your conversation history. You can also choose to delete your conversation history with Grok in this menu.
There are plenty of other AI tools that you may need to do this for if you’re a habitual user. We’ll continue to add to them here over time. The settings may be a bit obscured and hard to find, but if the option is there, you’ll typically find it in the data or privacy settings of the selected chatbot.
Read the full article here