Please look at the AI/algorithm that checks for phishing in forms. Creating a quiz for 13 years old students to take a question was added:
"what are the features of a strong password?"
as this part of our curriculum. this question was flagged as phishing (although the message says it is requesting personal data) when this question is not doing that at all. removing the question removed the notice.
while I appreciate a level of protection, and that it can be disabled, surely this feature can be improved to analyse the whole text of the question to allow the feature to remain, but not flag misidentified elements.