Where does it say that? https://chatgpt.com/#settings/Security merely says

Adds the highest level of account security by requiring stronger sign-in methods and applying stricter protections to help prevent unauthorized access.

That is not about model training.