Does ChatGPT train on my chats by default?

As of 2025, OpenAI states that for personal users on Free, Plus, or Pro plans, ChatGPT may use conversations to train and improve its models unless users explicitly opt out. For Enterprise, Team, and API users, OpenAI does not use customer inputs or outputs for model training by default. The Consumer Privacy at OpenAI page confirms: “We do not train on API, ChatGPT Enterprise, and ChatGPT Team customer data by default.”

The Data Controls FAQ explains that users can disable training by turning off the “Improve the model for everyone” toggle under Settings → Data Controls. The OpenAI Privacy Policy outlines the differences between OpenAI’s consumer and business/API services. In effect, personal users’ chats may be used for training unless opted out, while Enterprise, Team, and API users’ data remains excluded from model training by default.

How long does ChatGPT keep my chats/logs?

As of 2025, OpenAI clarifies that data retention depends on your account type and settings. For personal accounts, chats remain until you delete them. Once deleted, they are scheduled for removal within 30 days. Temporary Chats are automatically deleted after 30 days, though this timeline may be overridden by a legal hold, allowing deleted or Temporary Chats to be retained indefinitely during ongoing litigation.

API data is retained for 30 days by default or not at all if you request Zero Data Retention. For Enterprise and Team accounts, retention duration is set by workspace administrators, and deleted data is removed within 30 days unless legally required otherwise.

Deleted conversations are “scheduled for permanent deletion from OpenAI systems within 30 days.” as per the Retention Policy. Temporary Chats are automatically deleted after 30 days and are not used for model training as per Data Controls. API customers may request approval for Zero Data Retention or Modified Abuse Monitoring controls to configure how their data is retained, as stated in the Data Guide. By default, abuse monitoring logs are stored for up to 30 days. Personal accounts follow a 30-day deletion window, Temporary Chats auto-delete after 30 days, and Enterprise or API users follow administrator- or policy-defined retention settings.

How do I opt out of model training and still keep history?

You can turn off model training while keeping your chat history by disabling the “Improve the model for everyone” option in your account settings. This prevents new conversations from being used for training while keeping your previous chats accessible unless you delete them.

To adjust this setting, go to Profile → Settings → Data Controls, then toggle off “Improve the model for everyone.” This control applies only to future conversations- data already used for training cannot be removed. Your chat history remains visible in the sidebar unless you manually delete it as per Data Controls. You must be logged in for the setting to persist. Enterprise and API users do not need to opt out, as their data is excluded from training by default.

What’s different for Enterprise/Work/Gov vs personal plans?

There are distinct differences between personal and business plans regarding data handling, ownership, and training. Business users receive stronger privacy protections, no model training by default, and greater control over data retention. Personal users must manually opt out to disable model training and have more limited control over how long their data is retained.

OpenAI’s Enterprise Privacy documentation states: “We do not train our models on your business data by default. You own your inputs and outputs.” In contrast, the Consumer Privacy documentation notes that personal plan data may be used to improve the model unless users opt out.

Model training for personal accounts is enabled by default and requires opt-out, while for Enterprise and API plans, it is disabled by default. In terms of data ownership, personal users grant OpenAI a licence to operate the service, whereas Enterprise and API users retain ownership where legally permitted. Enterprise customers can also configure custom retention durations aligned with their organizational policies.

Where are retention settings and export/delete controls?

Personal users can manage data retention and export settings under Settings → Data Controls. Deleted chats are removed from OpenAI systems within 30 days, and users can export their data through the Privacy Portal.

Enterprise and Team users may have additional administrator-level retention and export options. Data can be exported via the Privacy Portal, and manually deleted chats are scheduled for removal from OpenAI systems within 30 days. ChatGPT Team and Enterprise accounts may include organization-administered retention and export settings.

Certain legal or technical exceptions may affect deletion timelines, and there is currently no auto-delete schedule (for example, 3, 18, or 36 months) available for personal users.

Is API usage treated differently from the app?

Yes,  OpenAI API usage follows different policies than ChatGPT web or app interactions. By default, API data is not used for training, and customers can request Zero Data Retention configurations, making it suitable for privacy-sensitive or regulated use cases. OpenAI states: “OpenAI does not use your API data to train our models unless you explicitly opt in.”

When Zero Data Retention is enabled, prompts and completions are discarded immediately after processing. In contrast, the consumer-facing ChatGPT app requires users to manually opt out to prevent model training. API usage operates under stricter data-handling standards, while Enterprise and Team environments are also excluded from model training and include enhanced audit and governance controls for greater administrative oversight.