Your data · Anthropic
How to stop Claude training on your chats
thecontinuitylayer.com/your-data/claude/turn-off-trainingWhether Claude has a switch that keeps your chats out of model training, and where it is.
Yes
Docs grade · You can stop your chats training its models
3 November 2026
Checked on Anthropic's pages
Read from published docs, not a test we ran
Report a change →Yes, according to Anthropic's own pages.
Anthropic's privacy page describes a Help Improve our AI models toggle under Settings, Privacy that Free, Pro and Max users can switch off themselves.
Anthropic's own pageChecked 3 November 2026
https://privacy.claude.com/en/articles/12109829How to do it in Claude
- 1.Open Claude on the web, the desktop app or the phone app.
- 2.Click your name to open the menu, then choose Settings.
- 3.Go to Privacy.
- 4.Turn off Help improve our AI models.
Turning it off stops future training on your past and new chats. Data already in a training run or an already-trained model stays there.
Anthropic's own pagesChecked 24 October 2026
The test: You can stop your chats training its models
Look in settings and on the privacy page for a control that stops your conversations being used to train or improve models.
- ✓ Yes
- A setting you can switch yourself, on consumer plans.
- ½ Partly
- Only by request or a form, or only on some plans, or off by default with no control named.
- ✕ No
- No way described, or chats are used with no opt-out.
A grade of what Anthropic's own page says, never a test we ran. Scores are never for sale. How the Scorecard works.