@ScottSeidman If what's reported here is wrong, I can't say anything at all. If someone wants to ask questions based on wrong content they're bound to get wrong answers.

The problem here goes beyond just the wrong interpretation, but into the wrong instructions for how to correct the problem, and the resulting approach indicated by OP which involves trying to get access to a detection software so that they can tweak the thing until it "passes". This is not a correct approach for either AI-generated content or more classical plagiarism