Protect Your Data: How to Stop ChatGPT, Gemini, Claude, and Meta AI from Training on It

Any query you input into a chatbot might be used to enhance its algorithms by default. While major players prefer to keep quiet about this practice, it raises significant privacy concerns for both individual users and businesses. Fortunately, you can disable this data collection in each tool’s settings. Here’s how to stop major AI tools from training on your content.

Preventing ChatGPT from Using Your Data

ChatGPT states that data collection, which is enabled by default, aims to “make the AI more accurate and efficient”. This applies only to free, Plus, and Pro accounts. Team and Enterprise plans, as well as API usage, are exempt.

However, you can choose not to allow your interactions to be used for training. The steps are the same whether on a computer or a mobile device:

• Open the profile menu (icon in the upper right corner on a computer, upper left on mobile).
• Go to Settings, then Data Management.
• In the section Improve the model for everyone, turn off the option.

Once this option is unchecked, your conversations will no longer be used to train ChatGPT. You can change this setting back at any time.

Preventing Gemini from Using Your Data

Google informs us: “By keeping your activity, you can resume your conversations at any time and help improve Google’s services, including its AI models. If this setting is turned off, Google still keeps the conversations for 72 hours to respond to you and ensure Gemini’s security.”

To turn off this feature, open Gemini and navigate to the Activity tab. Then click on Disable and delete activity to prevent the storage and use of your conversations. On mobile, these settings are accessible from the user profile icon, under the section Activity in Gemini apps.

Disabling this feature will delete the history and prevent recording of future conversations. You will not be able to retrieve your past exchanges.

For those who are particularly motivated, you can also submit a data processing objection request through Gemini’s privacy policy. The form is available at this link.

Preventing Claude from Using Your Data

In August 2025, Anthropic announced that Claude would now train on user conversations from the Free, Pro, and Max offers, unless users explicitly opt out. It only takes three clicks to stop your data from being used. Here’s what to do:

• Click on your profile icon to open the Settings.
• Go to the Privacy tab.
• Uncheck the box Help improve Claude.

Preventing Perplexity from Using Your Data

“Data retention allows Perplexity to use your searches to enhance AI models,” we are informed by the AI search engine. If you dislike this practice, follow these steps:

• Click on your profile icon.
• Select Preferences.
• Turn off the option titled AI Data Retention.

Preventing Le Chat from Using Your Data

As a French and sovereign tool, perhaps Le Chat is more scrupulous when it comes to using its users’ data? Unfortunately, that is not the case: Mistral has adopted a policy similar to that of ChatGPT, Gemini, and others. The procedure to oppose this practice is also similar:

• At the top left, click on your profile.
• Choose Settings.
• Go to Preferences.
• Disable the option named Allow the use of your interactions to train our models.

Preventing Copilot from Using Your Data

Copilot distinguishes two types of training: text-based, which is enabled by default, and voice-based, which is disabled by default. If you wish to refuse all training on your content, simply follow these steps:

• Click on your profile icon.
• Click on your name.
• Select Privacy.
• Disable Model learning on text (voice should already be disabled).

Preventing DeepSeek from Using Your Data

Similar to ChatGPT, the data usage option on DeepSeek is elegantly titled “Improve the model for everyone”. You can disable it by clicking on your profile, then on Settings and finally Data.

Preventing Meta AI from Using Your Data

Meta uses public posts from Instagram, Facebook, and Messenger to train its AI models, including Meta AI. Your comments, captions, and posts “visible to everyone” may therefore be involved. Private messages and data from minors are excluded. To refuse this usage, you must fill out a form accessible from the settings of your Meta accounts:

• Open the Settings of Instagram or Facebook, then the Privacy Center.
• Click on the link in the word “oppose”.
• Fill out the form with your email address and submit.

Once the form is sent, Meta will no longer be able to use your public content to train its models.

Preventing Grok from Using Your Data

If you’re savvy, you might be using Grok, the AI that Elon Musk struggles to mold in his image. This chatbot also uses data from your conversations. To prevent this, we recommend going through X:

• Press on More, then Settings and Privacy.
• Click on Privacy and Security.
• Select Grok and third-party collaborators.
• Uncheck Allow the use of my public data as well as my interactions, inputs, and results obtained with Grok and xAI for training and adjustment purposes.