Who Can See Your ChatGPT History?

Short answer: OpenAI can see your ChatGPT conversations, because they are stored on its servers. By default, a consumer account’s chats may be used to train OpenAI’s models unless you turn that off. Your employer or school may be able to access them if you use a managed workspace, and legal or security obligations can require OpenAI to keep data even after you delete it. Three controls reduce some of this exposure: the model-training setting, Temporary Chat, and chat deletion. None of them is a guarantee of zero retention.

What does OpenAI keep when you chat with ChatGPT?

Regular and archived chats stay in your account until you delete them or a workspace retention policy removes them, according to OpenAI’s chat and file retention page. Archiving is not deletion: an archived chat stays in your account and is treated like a regular chat.

When you delete a chat, ChatGPT removes it from your view immediately. OpenAI then schedules it for permanent deletion from its systems within 30 days. The exceptions are chats that were already de-identified and disconnected from your account, and chats OpenAI must keep longer for security or legal obligations (OpenAI: deleting and archiving chats). Deleted chats cannot be restored, even by OpenAI Support.

Two details people often miss: files saved separately in Library are not removed when you delete the chat that used them, and a chat that used a connected app keeps that app data inside the conversation until the chat is deleted.

Are your ChatGPT chats used to train AI models?

That depends on your plan and one setting. On personal accounts, the Improve the model for everyone setting controls whether your new conversations help train OpenAI models. To turn it off on the web, open your account menu, select Settings, then Data controls, then switch off Improve the model for everyone and select Done. On iOS and Android, open the sidebar, tap your profile icon, then Settings and Data controls (OpenAI: data controls in ChatGPT).

Turning the setting off does not delete or hide your saved chats; they remain in your history. It also applies only to new conversations, and if you submit feedback on a response, the related conversation can be used under OpenAI’s policy. For business products, OpenAI says it does not train on ChatGPT Business, Enterprise, Edu or API data by default (Enterprise privacy at OpenAI).

What does Temporary Chat change?

A temporary chat does not appear in your history, does not create or update memories, and is not used to improve OpenAI models while it stays temporary. OpenAI may still keep a copy for up to 30 days for safety purposes (OpenAI: Temporary chat). If you save a temporary chat, it becomes a regular chat and follows your normal training and retention settings.

So Temporary Chat reduces what is stored about you, but it is not zero retention. If you use a custom GPT with actions, data sent to a third party through an action is governed by that third party’s privacy policy, and it may keep the data longer than 30 days.

Can your employer or school see your ChatGPT chats?

If you use ChatGPT through a managed workspace, yes, the organization can set retention, compliance and access policies, and the OpenAI page notes that authorized compliance tools may still reach eligible workspace data. OpenAI says Enterprise, Edu and Healthcare workspace owners control how long data is retained (Enterprise privacy at OpenAI). For eligible Enterprise customers, temporary chats are available through the Compliance API for 30 days.

Separately, what an employer can see on a work device or network depends on how the organization manages that device and on its policies, not on OpenAI’s. Only that organization can tell you what it monitors; check its acceptable-use policy. The safest habit is to keep personal and sensitive topics out of work accounts and work devices.

Can the government or the police read your chats?

OpenAI’s retention page states that it may retain data longer when required for security or legal obligations. Deleting a chat removes it from your account and schedules it for deletion within 30 days, but it does not override a legal obligation to preserve data. If this matters to you, do not put the information into a chatbot in the first place.

How do other AI chatbots compare?

• Google Gemini: Google’s help page says Gemini Apps Activity auto-deletes after 18 months by default, and you can change that to 3 months, 36 months or indefinite. Chats reviewed by human reviewers, and related data, are kept for up to three years, disconnected from your Google Account, and are not deleted when you delete your activity. Temporary chats and chats with Keep Activity off are kept for 72 hours; Google says they are not used to train its models, though feedback you submit can be handled differently (Google: Gemini Apps privacy hub).
• Anthropic Claude: Anthropic lets you choose whether your chats are used to improve Claude. Deleted conversations are removed from your history immediately and from back-end storage within 30 days. If you allow training use, Anthropic may keep de-identified data for up to five years. Incognito chats are not used to improve Claude, though they are not zero-retention: Anthropic notes exceptions for safety, legal and feedback reasons (Anthropic: how long do you store my data?, Anthropic: is my data used for model training?).
• Apps built on top of these models: The app developer’s own policy applies, and it can differ from the model provider’s. Read it before you paste anything sensitive.

These policies change often. Check the linked pages for the current wording before you rely on a specific number.

How can you keep your AI chats private?

1. Turn off model training where the product offers it, in the settings of every chatbot you use (for ChatGPT: Settings, Data controls, Improve the model for everyone).
2. Use Temporary Chat or Incognito for one-off sensitive questions.
3. Delete chats you no longer need, and delete saved files in Library separately.
4. Leave out identifying details such as names, account numbers and health information.
5. Keep work and personal use apart. Do not use a work account for personal topics.
6. Consider running a model locally for truly sensitive material. When used in local mode, tools such as Ollama keep prompts on your own computer; its cloud features are a separate choice, so check which mode you are using.

Bottom line

Treat a cloud chatbot like a service that stores what you type. You control the training setting, deletion and Temporary Chat; your organization controls what happens in a managed workspace; and legal or security obligations can override deletion. Review the settings once, and re-check the linked policy pages when they change.