OpenAI's ChatGPT privacy policy says it collects your account details, your prompts and uploads, and technical data about your device and use. It may use your content to train its models unless you opt out, and it keeps data for different periods depending on the type and your settings. The US version was last updated on 10 September 2026. This page explains it and is not legal advice.
01Which ChatGPT privacy policy applies to you?
OpenAI publishes more than one version. The US privacy policy points readers in the European Economic Area, the United Kingdom and Switzerland to a separate version. Who is responsible for your data also differs: OpenAI Ireland Limited for people in the EEA or Switzerland, and OpenAI OpCo, LLC anywhere else. We read the US version only.
The policy does not cover content OpenAI handles for business customers such as API users; customer agreements govern that. It sends readers to a separate page on how your data is used to improve models for training.
02What does the policy say OpenAI collects?
Section 1 sorts data into what you provide and what OpenAI receives from your use.
- Account information: name, contact details, login details, date of birth, payment information and transaction history.
- User content: prompts and uploads such as files, images, audio, video and data from connected services.
- Log, usage and device data: IP address, browser, request times, features used, feedback on replies and device identifiers.
- Location: a general area from your IP address, and precise location only if you allow it.
- Contacts, if you choose to connect your address book.
- Other sources: security and safety partners, marketing vendors, advertisers, and public web pages used to build models. For Free and Go users, ad history and interests.
03What does OpenAI say it does with that data?
Section 2 lists the purposes: running and improving the service, research, personalising your experience, personalising ads for Free and Go users, contacting you, marketing, preventing fraud and misuse, and complying with the law. It also lists estimating your age to give an age-appropriate experience.
On training, the policy says OpenAI may use your content to train the models behind ChatGPT, and points to an opt-out. It also says it makes de-identified or aggregated data and uses it for the same purposes without trying to identify you again, unless the law requires it.
Section 3 says data can go to vendors and service providers, affiliates, marketing partners, authorities when the law requires or OpenAI believes a rule was broken, parties in a business sale, administrators of a business account you joined and, for linked teen accounts, a parent or guardian.
04What does it say about keeping data, your rights and children?
A policy can be overridden by a court. OpenAI says a 2025 legal order to keep consumer content set aside its normal 30-day practice until it ended on 26 September 2025.
- Deleting: if you delete chats, memories or your account, OpenAI says it removes the data within 30 days, unless it must keep it for security, safety or legal reasons or it was already de-identified.
- Temporary Chats are deleted automatically within 30 days, with the same exceptions.
- Longer retention: banned content or accounts, legal demands such as a subpoena, payment records and a record of your erasure request.
- Rights: depending on where you live, you may be able to access, correct, delete, restrict or move your data, and to complain to your data protection authority.
- Children: the service is not meant for under-13s, and people under 18 need permission from a parent or guardian.
05What does the policy not tell you?
A policy summarises practice; it does not prove it. These are gaps we noticed.
- It gives categories and a 30-day deletion window, not a retention period for every kind of data.
- It lists types of vendors, not their names.
- The pages we read do not explain how personal information is reduced in training data, or what happens to training already done if you opt out later.
- One reason for sharing with authorities is OpenAI's own judgment that a rule has been broken.
The policy itself says no online transmission is fully secure and asks users to take care over what they provide.
06How do you act on your rights?
The policy points to account tools first and a request portal second.
- In ChatGPT, open Settings and Data Controls to export, delete or control training. Labels are as OpenAI's help article describes them in October 2026, and menus change.
- For requests the app cannot handle, use OpenAI's privacy portal or the data request email address given in the policy.
- If you are unhappy with the answer, you can complain to your local data protection authority. For legal advice, ask a lawyer.
?What is not known
- OpenAI updates this policy often; always check the date at the top of the page.
- We did not read the EEA, UK and Switzerland version, so differences in that version are not covered.
- Regulators have not settled how privacy law applies to training chatbots on personal data. An Italian fine against OpenAI was set aside by a court in March 2026, and Ireland's regulator is now OpenAI's lead authority in the EU.
QQuestions people ask
Is the policy different for ChatGPT Enterprise or the API?
Yes. The policy says it does not cover content handled for business customers, which falls under customer agreements, and OpenAI says it does not train on business inputs or outputs by default.
Does the policy say OpenAI sells personal data?
Its US state disclosures say it does not sell personal data, but it may share limited data with select marketing partners and offers a marketing privacy control to opt out.
Can a teenager use ChatGPT?
The policy says people under 18 need a parent or guardian's permission, and teens and parents can link accounts so the parent can manage some settings. Under-13s are outside the service.
Last checked October 24, 2026. Rules and settings change, so open the source to see the current version.
- OpenAI US privacy policy OpenAI The policy dated 10 September 2026; every statement about what it says comes from this page.
- How your data is used to improve model performance OpenAI The companion page on training and opt-outs, updated 13 March 2026.
- Response to The New York Times data demands OpenAI OpenAI's account of a 2025 legal retention order and when it ended.
- Data Controls FAQ OpenAI Help Center Help article on Data Controls; read through search excerpts because the page blocked our reader.
- Garante press release on ChatGPT Garante per la protezione dei dati personali (Italy) Used for the status of the Italian fine and the move of the case to Ireland.
General information, not legal, medical, financial or investment advice. We are not affiliated with any company, agency or standards body named here. See our editorial policy and corrections, or tell us about a mistake at [email protected].