Right now it’s only possible to manage AI recognition by going to the already linked images to this person and then marking the incorrect ones.
I have some events where a person isn’t correctly recognized as the listed family member and would like to add such footage to the person for training purposes