Training models on documents i have uploaded

If i create a custom GPT and give that GPT access to only files that i upload rather than the generic ChatGPT dataset will OpenAI train it’s models on those documents?

I am wondering this because the documents will include semi-sensitive data.

Thanks in advance to anyone who can answer this querie.

yes, it can but depending on the nature of the content u may or may not have trouble depending on the API and the instructions u give it as long as it don’t violate the polices ur good