[throwaway22032]

Here's a human review process for you:

1) Check whether the output of the machine learning model outputs Yes or No.

2) If yes, ban.

3) If no, no ban.

This is not a human review process. The review process is algorithmic, the human is only involved to relay the result.