Stop Claude From Training on Your Data: Essential Privacy Tips Revealed!

In a statement released on Thursday, August 28, 2025, Anthropic declared that its AI models would soon be trained using data from Claude users, specifically their conversations and “coding sessions”, unless they choose to opt out.

Update Affects Both Free and Some Paid Subscribers

The updated terms of use and privacy policy will apply to free users, as well as subscribers to the Pro and Max plans. “These do not apply to services covered by our commercial terms, including Claude for Work, Claude Gov, Claude for Education, or API usage, for instance through third parties like Amazon Bedrock and Google Cloud Vertex AI,” the statement explains.

However, only new conversations and coding sessions with Claude will be used for training purposes, Anthropic clarifies. This decision is driven by the desire to enhance their product while also boosting “protections against malicious uses”. Additionally, the company has extended the retention period for chat data to five years, previously 30 days. “If you choose not to provide your data for training purposes, the retention period will remain at 30 days,” Anthropic adds.

How to Opt Out of Your Conversations with Claude Being Used?

Claude users have until September 28, 2025, to consent to the use of their data. Since this Thursday, Anthropic has implemented a pop-up window (see image at the top of the article) that appears upon relogging in, which allows users to “review these updates” and manage their settings. To prevent Anthropic from using your data, simply uncheck the option You can help improve Claude.

This can also be disabled through the settings by following these steps:

• Click on your profile name,
• Select Privacy,
• Uncheck the Help Improve Claude option.

New users will need to decide whether to allow their data to be used for training purposes when they create their account. This decision is not final and can be changed at any time by following the above procedure.