ChatGPT Data Privacy: What Happens to What You Type
ChatGPT data privacy comes down to one question with a tier-dependent answer: what does OpenAI do with what you type? The gap between a personal free account and a business tier is the entire privacy story, and most concern about ChatGPT at work disappears once a team is on the right tier and knows the rules. Here is what actually happens to your data on each.
Free and Plus: training is on by default
On consumer ChatGPT (the free tier and Plus), OpenAI may use your conversations to improve its models unless you turn off 'Improve the model for everyone' in data controls. Most users never change that setting. So work data pasted into a personal free or Plus account can flow into model training. This is the classic shadow-AI exposure, and it is why personal accounts are not appropriate for company data.
Team, Enterprise and the API: not trained on by default
OpenAI does not train on inputs or outputs from ChatGPT Team, ChatGPT Enterprise, or the API platform by default. These tiers also add admin controls, SSO on Enterprise, and a Data Processing Agreement for GDPR. For any work involving client data, personal data, or confidential plans, ChatGPT Team is the minimum sensible tier, and Enterprise adds SSO and configurable retention on top.
Retention: how long chats stick around
Consumer chats are retained until you delete them, with deleted conversations typically purged around 30 days later. ChatGPT Enterprise and Team offer customer-configurable retention. The API defaults to roughly 30 days of logging, with Zero Data Retention available for eligible endpoints. If retention matters for your use case, the business and API tiers are where you get control over it.
What this means for your team
The practical ChatGPT data privacy policy is short: use the company ChatGPT Team or Enterprise workspace, never personal accounts, for anything work-related; do not enter secrets, customer data, or confidential plans into any consumer AI tool; and record the approved tier in your AI usage policy so it is a rule rather than a hope. ModelCharter's directory has ChatGPT's full data-handling profile, sourced from OpenAI's own policies, and the policy generator turns the tier decision into a signed team rule.