Privacy/level of exposure - CustomGPT


So I have a question regarding privacy in CustomGPTs. I read another thread from another user with a similar question but it does not quite answer my question. Essentially in CustomGPT, there is a button called “Use conversation data in your GPT to improve our models” that can be unchecked once a file is uploaded under “configure” then “knowledge”.

However, I was wondering if there was a similar function generally, if I decide to upload a file to my CustomGPT once it is published as private. That means that my file is not uploaded in the “knowledge” section but inside the actual chat log. I don’t want OpenAI to use such info in its training.

Same question regarding anything I type and feed a published private CustomGPT. Any way of stopping such data that I upload from being used in OpenAI training subsequently?

I read on another thread that a ChatEnterprise account may have this layer of privacy, can anyone confirm (I am simply on ChatGPT Plus at the moment)? I know that I can achieve this privacy using the API but I would simply prefer a prebuilt solution which is faster to deploy if possible.

Thanks for the clarification!

For privacy “assurance” you need to upgrade to a ChatGPT Team account.

1 Like

Thanks for the reply! Before doing so, would you be able to elaborate on the privacy guarantees it would offer, ie, would the data I send to the GPTs I make using my enterprise account be used in any way to train OpenAI’s models?

Thanks again!


You can read up on the Team subscription here:

There are some caveats when migrating from Plus to Team and eventually back but if it’s privacy and GPT usage you are after, then this is the available option.