Your team's "No training" policy now covers ChatGPT too
Starting October 15, your team's "No training-data sharing" policy also checks ChatGPT's "Improve the model for everyone" setting.
What is it?
When "Improve the model for everyone" is on, what you enter in ChatGPT (questions, pasted documents or code, uploaded files) may be used to train OpenAI's models. On personal plans (Free, Go, Plus, Pro) your chats can be used for training unless you turn this off. Business and Enterprise workspaces are not used for training by default.
If you use ChatGPT for work: if you put customer data, internal documents or source code into ChatGPT, we recommend turning this setting off.
Why did we add it?
Teams that chose "No training" want their work conversations kept out of training. Yet about 68% of personal ChatGPT accounts we see have this setting on. So we now check it and notify you the same way we do for Claude.
What changes?
- In "No training" teams, members whose personal ChatGPT account has this setting on get the same notice email as for Claude — automatic notices at most once every 7 days.
- Team admins get a copy with the list of those members.
- Personal accounts a member has hidden from the team are not checked.
Where do I see it?
- Team dashboard: the training-data KPI, member badges, and Settings → Training data policy → violators list
- Extension popup: a "Sharing training data" row appears when the ChatGPT account you are viewing has it on
How do I turn it off?
- ChatGPT → account menu → Settings → Data controls → "Improve the model for everyone" → off → Done
- Teams that do not need the policy can switch it to "Allow" in Team dashboard → Settings — no emails are sent then (applies to Claude and ChatGPT).
Learn more: OpenAI Help — Data controls in ChatGPT