Do ChatGPT and Claude use your chats to train their AI?
Both companies’ privacy policies say they may use what you put into your chats to train or improve their AI models. OpenAI’s says that training is subject to your control; Anthropic’s says you can opt out in your account settings, and when it will still use your chats for model improvement after you do.
Written October 26, 2026 from the policies as Plainlyst read them that day.
What this covers
This guide covers the privacy policies for personal ChatGPT and Claude use: for ChatGPT, the page OpenAI titles “US privacy policy”. Both say they do not apply to content the company processes for its business customers — OpenAI names its API and Anthropic its Enterprise accounts — and both say their use of that content is governed by their customer agreements, not by these policies.
ChatGPT (OpenAI)
OpenAI’s privacy policy says it may use the content you provide to improve its services, including to train the models that power ChatGPT, and that training on your content is subject to your control.
we may use Content you provide us to improve our Services, for example to train the models that power ChatGPT
including using your Content to train our models (subject to your control)
From the OpenAI Privacy Policy (U.S.) (Updated: September 10, 2026). Read OpenAI Privacy Policy (U.S.) (opens in a new tab)
The date is from the text of OpenAI’s policy page as fetched on October 26, 2026; Plainlyst kept that text, not a copy of the page. Plainlyst’s ChatGPT report was read from the May 18, 2026 edition; the passages quoted here appear word for word in both.
In Plainlyst’s Privacy Tune-Up: “Stop your conversations being used to train OpenAI's models” and “Use a Temporary Chat for one conversation”.
Claude (Anthropic)
Anthropic’s privacy policy says it may use your inputs and outputs to train and improve its models unless you opt out in your account settings, and it names two cases where it will still use them after you opt out.
We may use your Inputs and Outputs to train and improve Anthropic AI models, unless you opt out through your account settings.
Even if you opt-out, we will use Inputs and Outputs for model improvement when: (i) your conversations are flagged for safety review to improve our ability to detect harmful content, enforce our policies, or advance AI safety research, or (ii) you've explicitly reported the materials to us […]
So opting out does not cover conversations flagged for safety review, or materials you have explicitly reported to it. The full passage, with its example, is on Anthropic’s page.
From the Anthropic Privacy Policy (Effective Sep 10, 2026). Read Anthropic Privacy Policy (opens in a new tab)
The date is from Anthropic’s policy page as fetched on October 26, 2026. Plainlyst’s Claude report was read from the July 8, 2026 edition; the passages quoted here appear word for word in both.
In Plainlyst’s Privacy Tune-Up: “Turn off using your chats to improve Claude” and “Use an Incognito chat”.
Where the settings are listed
Plainlyst’s Privacy Tune-Up lists these settings under each company, and each company’s report links to the first of them from its training finding. Where to find each setting and what it changes are in Premium.
Worth knowing
- These passages say what each company may do. They do not show what happened to any particular conversation.
- Policies change. Each section gives the date its policy states; Plainlyst read both on October 26, 2026.
About Plainlyst
Plainlyst
Operated by JC RESERVE LLC
Official sources only. Every finding quotes the company’s own policy, and every record names the document behind it. We do not grade anybody.