Many people wonder whether the conversations they have with Claude — including personal thoughts or work documents — end up as training data for AI. The short answer is: it depends on your plan and your privacy settings. This article walks through what Anthropic's official privacy policy and help documentation say about when your data is used for training, and what you can do about it.
Key Terms, Briefly Explained
- Training Data: Example conversations and text used to help an AI model generate better responses over time.
- Privacy Policy: The official document from Anthropic that describes what data is collected, stored, and how it may be used.
- Conversation History: The running list of messages in your current session that lets Claude keep track of context during a conversation.
- Data Retention: How long your conversation data is stored on Anthropic's servers.
The Core Principle: How Does Claude Handle Conversations?
According to Anthropic's official privacy policy, conversations on claude.com may be processed to improve service quality, safety research, and related purposes. This does not automatically mean your conversation is fed directly into model training as-is. Whether your data is used for training depends specifically on your plan and privacy settings.
How Plans Differ
Free, Pro, and Max Plans (Personal Subscriptions)
On personal plans, conversations may be used to improve Claude's models by default. However, Anthropic provides a way to opt out. Turning off the relevant toggle in your privacy settings means your subsequent conversations will not be used for training. Keep in mind: disabling this setting also turns off conversation history, so Claude won't remember previous messages across sessions — every conversation starts fresh.
Team and Enterprise Plans (Organization Subscriptions)
According to official documentation, Team and Enterprise plans are set up so that conversations are not used for model training by default. These plans are designed with organizations in mind, where stronger data protection is expected.
Direct API Access
According to official documentation, conversations sent via the API directly are not used for training by default.
How to Opt Out of Training Data Use
If you're on a Free, Pro, or Max plan and want to opt out, here are the general steps:
- Log in to claude.com and open Settings — typically found in the lower left or via your profile area.
- Look for a Privacy or Data & Privacy section. The exact label may vary with interface updates, so confirm what you see on screen.
- Find the toggle for something like Improve Claude for Everyone and switch it Off.
- Once saved, the change applies to conversations you start going forward.
As noted above, turning this off also disables conversation history. Claude will no longer carry context between separate sessions. Make sure that trade-off works for you before switching.
Frequently Asked Questions
Q: Does changing my settings delete conversations I've already had?
Official documentation does not explicitly state whether the setting change applies retroactively to already-stored conversations. To request deletion of past data, you can use the conversation history deletion feature in settings, or submit a privacy request to Anthropic directly. For the current process, check support.claude.com.
Q: If Claude remembers my name or context, is that being used for training?
Claude's ability to remember your name or context within a conversation is about maintaining conversational flow within a single session. That alone does not mean the information is being used for model training. Training use is determined by the plan and settings conditions described above.
Q: What about third-party apps built on the Claude API?
When you use an external app or service that connects to Claude via the API, that service's own privacy policy also applies. Anthropic's policy covers its direct relationship with you, but the third-party developer may collect or handle data differently. Always review the terms of the specific app you are using.
Q: Is it safe to paste sensitive work documents into Claude?
Team and Enterprise plans are designed so that data is not used for training by default. That said, regardless of which plan you use, it is good practice not to include unnecessary personal data or confidential information belonging to others when using any AI service.
Summary
Whether Claude uses your conversations for training comes down to your plan and the privacy settings you control. Personal plan users can opt out in settings; Team and Enterprise users are protected by default. Since privacy policies can be updated at any time, we recommend checking the latest information directly at support.claude.com.