Short answer
All three major assistants let you limit training use of your chats and delete history: ChatGPT via Data Controls and Temporary Chat, Claude via the Model Improvement setting and incognito chats, Gemini via Keep Activity. Deleted chats can still linger for weeks, and settings change, so check each vendor's current help page.
Key takeaways
- Training opt-outs and history deletion are separate switches. Do both if you want less data kept.
- Temporary, incognito and Keep Activity off modes exist in ChatGPT, Claude and Gemini, but each still holds chats for a short safety window.
- Deleting a chat does not erase it everywhere at once: Claude describes back-end removal within 30 days, and ChatGPT keeps temporary chats up to 30 days.
- Google says reviewed Gemini chats are disconnected from your account and kept up to three years, and deleting your activity does not remove them.
- Menu names change often. Use the official help links in this article to confirm the current steps.
Two settings, not one
Every major assistant separates two questions. Is my chat used to improve the model? And how long is it stored? Opting out of the first rarely changes the second. This guide covers ChatGPT, Claude and Gemini for personal accounts as of 5 November 2026. Menus move, so confirm against the linked help pages.
ChatGPT (OpenAI)
Go to Settings, then Data controls. Switch off “Improve the model for everyone” so new conversations are not used for training. They still show in your history until you delete them.
Use Temporary Chat for anything sensitive. OpenAI’s help pages say these chats are not used for training, do not appear in history, and may be kept for up to 30 days for safety purposes. In Data controls you can also delete chats, and you can manage any shared links you created. Wording on OpenAI’s pages has varied between “deleted after 30 days” and “retained up to 30 days”, so read the current version.
Claude (Anthropic)
Claude’s Privacy settings include a “Model Improvement” toggle. Anthropic says chats may be used for training if you opt in, if a conversation is flagged for safety review, or if you explicitly join a program such as its Trusted Tester Program. Incognito chats are not used to improve Claude, even with Model Improvement on.
On retention, Anthropic’s retention page says deleted chats leave your history immediately and are removed from back-end storage within 30 days. If you opted in to training, de-identified chats may be kept up to five years. Feedback given with the thumbs up or down button is kept with the whole conversation for up to five years. Chats flagged for policy violations can be kept longer.
Gemini (Google)
Gemini’s control is “Keep Activity” in Gemini Apps Activity. By default activity auto-deletes after 18 months, and you can change that to 3 or 36 months, or delete manually. With Keep Activity off, or in a temporary chat, Google says chats are kept with your account for 72 hours and not used to train its models. Turning it off also stops some connected apps from working.
Two details matter. Some chats are reviewed by humans, and Google warns not to enter anything confidential you would not want a reviewer to see. Reviewed chats are disconnected from your account and kept up to three years, so deleting your activity afterwards does not remove them.
What settings cannot do
- They do not stop safety or abuse review.
- They do not recall data already used in a finished model. Anthropic notes data already in a trained model remains.
- They do not protect you if your account is taken over. Use a strong password and two-step verification.
- They do not control what you paste. See what not to tell AI chatbots.
A sensible routine
Turn off training, use temporary modes for private topics, delete old chats monthly, and remove shared links. If an AI account is part of a breach, our data breach guide explains what to do next, and our breach tracker lists recent incidents.
What you can do
- 01ChatGPT: open Settings, then Data controls, and switch off 'Improve the model for everyone'. Use Temporary Chat for sensitive questions.
- 02Claude: open Privacy settings and turn off 'Model Improvement'. Use incognito chats when you want nothing used for training.
- 03Gemini: open Gemini Apps Activity (myactivity.google.com/product/gemini), turn off 'Keep Activity', and set auto-delete to 3 months if you leave it on.
- 04Delete old conversations you no longer need, including ones with personal details, and remove any public share links.
- 05Set a reminder to recheck these settings every few months, or after a major app update.
Related reading: AI voice clone scams and the data breach tracker.
Frequently asked questions
Does opting out of training delete my chats?
No. In ChatGPT, turning off 'Improve the model for everyone' means new conversations are not used for training but still appear in your history. Deleting is a separate step.
How long does Claude keep deleted chats?
Anthropic says deleted chats leave your history immediately and are removed from back-end storage within 30 days. Chats flagged for policy violations can be kept longer.
Can people at Google read my Gemini chats?
Some chats are reviewed by human reviewers to improve services. Google says reviewed chats are disconnected from your account and kept up to three years, and advises against entering confidential information.
Are paid and work accounts different?
Often yes. Business, enterprise and education plans usually have separate terms, and training on your data is typically off by default. Ask your administrator rather than assuming.
Sources
- Is my data used for model training?, Anthropic Privacy Center, 2026-03-16
- How long do you store my data?, Anthropic Privacy Center, 2026-07-01
- Gemini Apps Activity and human review, Google Help, 2026-11-05
- Data Controls FAQ, OpenAI Help Center, 2026-11-05
This article reflects public information as of 2026-11-05. AI tools and rules change quickly; we update articles when new facts are published. Spot an error? See our editorial and corrections policy.