Does OpenAI train on my ChatGPT data?

By Hugh Donatello, LAN Service Group · Reviewed October 20, 2026

Short answer

It depends on the plan. On Free, Plus and Pro, OpenAI says it may use your content to train its models unless you turn off "Improve the model for everyone" in Data controls or opt out in its Privacy Portal. By default, OpenAI does not train on inputs or outputs from ChatGPT Business, Enterprise, Edu or the API. Retention, Temporary Chat and admin controls also differ by plan.

Training, opt-out and retention settings by plan

This table covers one question: what happens to the content of a chat or API call on each OpenAI plan. Every row cites the OpenAI page it comes from and the date we checked it. For features, identity controls and certifications, see our separate comparison of ChatGPT Business and Enterprise.

Does OpenAI train on your data? Settings by plan
Plan	Used to train OpenAI models by default?	How to keep content out of training	Temporary Chat	Retention after you delete	Who controls the settings	Source (checked on)	Source
ChatGPT Free	It may be. OpenAI says it may use content from its services for individuals to train its models.	Turn off "Improve the model for everyone" under Settings > Data controls, or select "Do not train on my content" in OpenAI's Privacy Portal. Either one is enough. Giving thumbs-up or thumbs-down feedback can still make that whole conversation eligible for training.	Not used to improve models while it stays temporary. It does not appear in chat history or update memories, and is deleted automatically within 30 days unless OpenAI must keep it for safety or legal reasons. Saving it turns it into a regular chat.	Deleted chats are removed from OpenAI's systems within 30 days, unless OpenAI must keep them longer (for example, for legal, safety or security reasons) or they were already de-identified and disassociated from your account for model training.	The individual user (for a linked teen account, a parent or guardian can manage this setting). If the account was opened with a work email, OpenAI may share that the account exists with the employer so it can be added to the employer's business account.	OpenAI Help Center, How your data is used to improve model performance; OpenAI Privacy Policy (checked October 19, 2026)	[1] [3]
ChatGPT Plus (and Pro)	It may be. OpenAI groups Free, Plus and Pro together as services for individuals. Paying for Plus or Pro does not change the training default.	Same as Free: the Data controls switch or the Privacy Portal. The feedback rule also applies.	Same as Free.	Same as Free.	The individual user, as on Free.	OpenAI Help Center, How your data is used to improve model performance; OpenAI Privacy Policy (checked October 19, 2026)	[1] [3]
ChatGPT Business	No. OpenAI says it does not use prompts, responses or files from Business to improve its models unless the organization has signed a separate agreement allowing it or has explicitly opted in to share data.	Off by default; confirm nobody has opted the workspace in to data sharing. Thumbs-up or thumbs-down feedback does not permit training on the conversation, and admins can turn response feedback on or off under Workspace settings > Allow response feedback.	OpenAI's help article describes Temporary Chat without limiting it to individual plans. For Business, OpenAI says deleted or unsaved conversations are removed within 30 days (exceptions as in the retention column). Workspace chats are already excluded from training by default.	Workspace admins can control how long data is kept. Deleted or unsaved conversations are removed within 30 days, unless longer retention is required by law or is reasonably necessary to protect OpenAI's services or a third party from harm.	Workspace admins. They can view, access, export and delete members' conversations in the workspace.	OpenAI Help Center, How your data is used to improve model performance; OpenAI, Enterprise privacy at OpenAI (checked October 19, 2026)	[1] [2]
ChatGPT Enterprise	No. The same no-training default, and the same exceptions (a separate agreement or an explicit opt-in to share data), apply to Enterprise and Edu.	Off by default; confirm nobody has opted the workspace in to data sharing. Feedback does not permit training, and admins control the response feedback setting.	OpenAI's Enterprise privacy page does not address Temporary Chat separately. Workspace chats are already excluded from training by default.	Workspace admins control how long data is kept. Deleted conversations are removed within 30 days, unless OpenAI is legally required to keep them.	The organization. Users see their own conversations, and admins can get an audit log of conversations and GPTs through the Enterprise Compliance API.	OpenAI Help Center, How your data is used to improve model performance; OpenAI, Enterprise privacy at OpenAI (checked October 19, 2026)	[1] [2]
OpenAI API	No. Since March 1, 2023, data sent to the API has not been used to train or improve OpenAI models unless the customer explicitly opts in to share data.	Sharing is off unless enabled. API organization owners can enable data sharing, which lets members share feedback and the current API Playground chat with OpenAI. Check that this has not been turned on without a decision.	Not applicable. Temporary Chat is a ChatGPT feature.	Abuse-monitoring logs, which may contain prompts and responses, are kept for up to 30 days on most endpoints (until deleted for Conversations and ChatKit threads), unless longer retention is required by law or is reasonably necessary to prevent harm. The Responses API stores response data for at least 30 days by default. Assistants API objects are kept until you delete them, then removed from OpenAI's servers 30 days after you delete them.	The API organization. Zero Data Retention and Modified Abuse Monitoring need OpenAI's prior approval and acceptance of additional requirements. Once approved, they are set under Settings > Organization > Data controls > Data Retention, for the whole organization or per project.	OpenAI Help Center, How your data is used to improve model performance; OpenAI API docs, Data controls in the OpenAI platform (checked October 19, 2026)	[1] [4]

Each row was checked against the OpenAI pages named in its last column on October 19, 2026. OpenAI updates these pages often (its help article on model training showed a recent update when we checked it), so confirm the current versions before relying on them in a contract, policy or vendor review.

Free, Plus and Pro: training is on until you turn it off

OpenAI treats Free, Plus and Pro as services for individuals, and on those plans it says it may use your content to train its models. Paying for Plus or Pro does not change that default. OpenAI's Privacy Policy also mentions a Go plan, which the help article does not list, so check that plan's Data controls directly. OpenAI says it takes steps to reduce the amount of personal information in training datasets before using user content, but that is not the same as keeping your content out.

There are two ways to opt out, and OpenAI says either one is enough for ChatGPT conversations: turn off "Improve the model for everyone" under Settings > Data controls, or choose "Do not train on my content" in OpenAI's Privacy Portal. OpenAI notes that Codex has its own separate setting for training on full environments, which the ChatGPT switch does not change.

Three details catch people out. First, the opt-out applies going forward: OpenAI says that after you opt out it will not use your new conversations. Second, feedback overrides it: on individual plans, if you give a thumbs-up or thumbs-down on a response, OpenAI may use the entire conversation behind that feedback to train its models, even after you have opted out. Third, deleting a chat may not reach everything: OpenAI's Privacy Policy says deleted personal data is removed within 30 days, except where it must be kept longer or has already been de-identified and disassociated from your account for model improvement.

Temporary Chat is the per-conversation option. OpenAI says temporary chats are not used to improve its models while they remain temporary, do not appear in your history and do not create or update memories. Its Privacy Policy says they are deleted automatically within 30 days unless they must be kept for safety or legal reasons. If you save a temporary chat, it becomes a regular chat and follows your account's normal settings.

Business and Enterprise: off by default, but admins still decide a lot

On ChatGPT Business, Enterprise and Edu, OpenAI says it does not use prompts, responses or files to improve its models unless the organization has signed a separate agreement allowing it or has explicitly opted in to share data. Unlike on individual plans, a thumbs-up or thumbs-down does not open the conversation to training. Admins can still decide whether staff can send response feedback at all, under Workspace settings > Allow response feedback in the Admin Console. OpenAI also says that, by default, it does not train on data that ChatGPT reaches through connected apps in these workspaces.

The no-training default is not the whole privacy picture. On Business, OpenAI says workspace admins can view, access, export and delete members' conversations. On Enterprise, users see their own conversations and admins can pull an audit log of conversations and GPTs through the Enterprise Compliance API. Either way, staff should know that their work chats are visible to the company, and your AI policy is the place to say so.

Retention is also an admin decision. OpenAI says Business admins can control how long data is kept and that Enterprise admins control it. On Business, deleted or unsaved conversations are removed within 30 days unless longer retention is required by law or is reasonably necessary to protect OpenAI's services or a third party from harm. On Enterprise, deleted conversations are removed within 30 days unless OpenAI is legally required to keep them. OpenAI notes that retention enables features like conversation history and that shorter retention periods may compromise the product experience, so choose a period that matches your records policy, not simply the shortest one.

Our comparison of ChatGPT Business and Enterprise covers the identity, audit and data-residency differences between the two plans. The point here is narrower: on both, training is off by default, and what is left to configure is feedback, retention and who can see what.

The API: a separate product with its own settings

The OpenAI API has its own data rules. OpenAI's platform documentation says that since March 1, 2023, data sent to the API has not been used to train or improve OpenAI models unless the customer explicitly opts in to share data. OpenAI's help center adds that API organization owners can enable data sharing, which lets members share feedback and the current API Playground chat to help improve models. Because sharing is an opt-in at the organization level, the setting to check sits in your API organization, not in anyone's ChatGPT account.

Not training on your data is not the same as not keeping it. By default, OpenAI generates abuse-monitoring logs for all API usage, and those logs, which may contain prompts and responses, are kept for up to 30 days on most endpoints unless longer retention is required by law or is reasonably necessary to prevent harm. Eligible customers can apply for Zero Data Retention or Modified Abuse Monitoring, which exclude customer content from those logs, subject to limitations OpenAI lists (Modified Abuse Monitoring, for example, can still log image and file inputs in rare cases). Both need OpenAI's prior approval and acceptance of additional requirements, and once approved they appear as a Data Retention tab under Settings > Organization > Data controls, where they can be set for the organization or project by project.

Some API features store data so they can work at all. OpenAI says the Responses API keeps response data for at least 30 days by default. Its endpoint table lists Conversations, ChatKit threads, assistants and threads among the features that keep application state until you delete it, with Assistants API objects removed 30 days after deletion. It also notes that features not eligible for Zero Data Retention may store data even when it is enabled. Whoever builds an internal tool on the API should write down which endpoints it uses and how it deletes what they store.

A short checklist for IT

Training is only one of the questions. For most businesses the bigger risk is staff using personal Free or Plus accounts for work, which our guide to whether employees can use ChatGPT with company data covers in full. Once a business plan is in place, these checks confirm the settings match what your policy says.

Personal accounts: if staff must use Free, Plus or Pro for anything work-related, require the "Improve the model for everyone" switch to be off, and remember that feedback can still send a conversation to training.
Response feedback: on Business and Enterprise, decide whether to allow it under Workspace settings > Allow response feedback, and record the decision.
Retention: set the workspace retention period to match your records policy, and note that deleted conversations can take up to 30 days to leave OpenAI's systems, or longer where OpenAI's stated exceptions apply.
Admin visibility: tell employees who can see their workspace conversations and that Enterprise conversations can reach an audit log.
API organization: confirm data sharing is off unless you chose to turn it on, list the endpoints that store data, and decide whether to apply for Zero Data Retention.
Review date: recheck these OpenAI pages every quarter, because OpenAI updates them often.

Common follow-up questions

Does OpenAI train on ChatGPT Plus conversations?

It may, unless you opt out. OpenAI groups Free, Plus and Pro as services for individuals and says it may use their content to train its models. Turn off "Improve the model for everyone" under Settings > Data controls, or choose "Do not train on my content" in OpenAI's Privacy Portal.

If I opt out, can a thumbs-up still send my chat to training?

On Free, Plus and Pro, yes. OpenAI says that if you give feedback on a response, it may use the entire conversation behind that feedback to train its models, even if you have opted out. On Business, Enterprise and Edu, feedback does not permit training on the conversation.

Are Temporary Chats deleted right away?

Not necessarily. OpenAI says temporary chats are not used to improve its models while they remain temporary and do not appear in your history, but its Privacy Policy says they are deleted automatically within 30 days, unless OpenAI must keep them for safety or legal reasons.

Does deleting a chat take it back out of training?

Not necessarily. OpenAI's Privacy Policy says deleted data is removed within 30 days, except where it must be kept longer or has already been de-identified and disassociated from your account for model improvement. Opting out covers new conversations, so opt out before you share anything sensitive.

Does OpenAI train on ChatGPT Business or Enterprise data?

Not by default. OpenAI says it does not use prompts, responses or files from Business, Enterprise or Edu to improve its models unless the organization has signed a separate agreement allowing it or has explicitly opted in to share data.

Is API data used for training?

Not by default. OpenAI says data sent to the API since March 1, 2023 is not used for training unless the customer explicitly opts in. API organization owners can enable data sharing, so check that setting. Abuse-monitoring logs are still kept for up to 30 days by default on most endpoints.

Need help with this?

LAN Service Group helps small and mid-size businesses set up company-controlled ChatGPT and API accounts, configure feedback, retention and data-sharing settings, and write the AI policy that tells staff what those settings mean.

Talk to LAN Service Group LAN Service Group AI services → (888) 281-7243