AIPolicyLog

Short answer · checked 2026-11-18

Does Claude train on your data?

Yes, on consumer Claude plans (Free, Pro, Max and Ultra) unless you opt out: Anthropic’s privacy policy says it may use your inputs and outputs to train its models; Claude for Work, the API and Claude Gov are not used for training by default.

Based on Anthropic’s own pages, checked 2026-11-18 (America/Toronto (UTC−5)). Full quotes for every plan: Anthropic / Claude policy by plan.

Trains by default?
Yes, unless you opt out (consumer)
Opt-out
Yes, Model Improvement setting
Business plans excluded?
Yes, not trained by default
Retention
Deleted chats 30 days; up to 5 years if allowed
Zero data retention
API, by arrangement

Personal and consumer accounts

Consumer inputs and outputs may be used for training unless you turn off Model Improvement in your Privacy Settings. Anthropic says turning it off means it will not use your previous or new chats or coding sessions for future model training. Even after opting out, Anthropic says conversations flagged for safety review, and material you explicitly report (for example through feedback), can still be used.

Business, enterprise and API

Commercial products (Team, Enterprise, API, third-party platforms and Claude Gov) are not used for training by default. An organization can opt in, for example through the Development Partner Program.

How to opt out

Applies to: Claude Free, Pro, Max and Ultra. Steps as documented by Anthropic (source, checked 2026-11-18).

  1. Open your Privacy Settings in Claude (claude.ai/settings/data-privacy-controls).
  2. Turn off Model Improvement. Anthropic says it then won’t use your previous or new chats or coding sessions for future model training (training runs already in progress are not undone).
  3. Avoid sending feedback on chats you don’t want used, because reported material can still be used after you opt out.

How long is your data kept?

Deleted consumer chats leave Anthropic’s back-end systems within 30 days. If you allow model improvement, de-identified data may be kept up to 5 years in training pipelines. Chats flagged for Usage Policy violations are kept up to 2 years. Commercial chats follow the organization’s retention setting.

Questions

Does Claude train on your data?

Yes, on consumer Claude plans (Free, Pro, Max and Ultra) unless you opt out: Anthropic’s privacy policy says it may use your inputs and outputs to train its models; Claude for Work, the API and Claude Gov are not used for training by default.

Can you stop Claude from using your data for training?

Claude Free, Pro, Max and Ultra: Open your Privacy Settings in Claude (claude.ai/settings/data-privacy-controls). Turn off Model Improvement. Anthropic says it then won’t use your previous or new chats or coding sessions for future model training (training runs already in progress are not undone). Avoid sending feedback on chats you don’t want used, because reported material can still be used after you opt out.

Does Claude train on business or API data?

Commercial products (Team, Enterprise, API, third-party platforms and Claude Gov) are not used for training by default. An organization can opt in, for example through the Development Partner Program.

How long does Claude keep your data?

Deleted consumer chats leave Anthropic’s back-end systems within 30 days. If you allow model improvement, de-identified data may be kept up to 5 years in training pipelines. Chats flagged for Usage Policy violations are kept up to 2 years. Commercial chats follow the organization’s retention setting.

Sources

Not legal advice. We restate the vendor’s own words; when a vendor doesn’t say something, we say so instead of guessing. How we compile this.

Compare

Get an email when Claude’s data policy changes

Free email alerts when Claude changes their data, training or retention terms. No spam. Unsubscribe anytime.