“Will my chats be used to train the AI?” is one of the most common questions Claude users ask. The answer: it depends on your account type and your own setting. On personal (consumer) plans, you choose whether to provide your data for model training, and that choice changes how long your data is retained. This guide covers how the training setting works, retention periods, differences by account type, and how to check your setting, based on Anthropic’s official announcement. (As of June 2026. Policy and menu locations may change — see the official announcement and your privacy settings for the latest.)
The training setting — you choose
With its 2025 consumer terms update, Anthropic made it so users on personal plans (Free, Pro, Max — including Claude Code used from those accounts) choose whether their chats are provided for model training. Turn it on and your subsequent new and resumed chats and coding sessions may be used; turn it off and they are not. The choice is not locked in — you can change it anytime in Privacy Settings, and changes apply to new or resumed chats going forward.
Retention — 5 years vs 30 days
Per the official announcement, turning the setting on extends retention to up to 5 years (including feedback you submit about responses), while turning it off keeps the existing 30-day retention. Some protections apply either way: deleted conversations are not used for future model training, Anthropic states it does not sell user data to third parties, and sensitive data is filtered or obfuscated with automated tools. Note that the privacy policy includes an exception allowing some conversations to be retained longer for trust and safety purposes (such as abuse prevention), so check the official policy documents for details.
Differences by account type
This training setting applies to personal (consumer) plans only. Per the official announcement, Claude for Work (Team and Enterprise), Claude for Education, Claude Gov, and Claude API usage (including via Amazon Bedrock and Google Vertex AI) are not in scope of this change and keep their existing protections. If you handle company or client data, the commercially-governed options are the better fit. For plan differences, see the pricing guide.
How to check your setting
In claude.ai, the model-training toggle lives under Settings → Privacy in your profile menu (names and locations can move with updates). Check your current state there and change it anytime. Good habits to pair with it:
- Regardless of the setting, avoid putting sensitive content (passwords, ID or card numbers, other people’s personal data, medical records) into any chat.
- Delete sensitive past conversations — deleted chats are not used for training.
- If you use the memory feature, see what gets stored in the memory guide.
For other data-related questions (language support, free tier, and more), see the FAQ.
The policy details here (opt-in training, 5-year/30-day retention, scope) reflect Anthropic’s official announcement on the consumer terms update and may change. For the latest, see the official announcement, the privacy policy, and your own account settings. This site is not an official Anthropic site.