The visitor (user) is scanned with a gun like device. A facial recognition software with a machine learning model trained on faces of gun users from youtube videos and selfies determine weather the person facial characteristics fit the pattern of either HIGH risk gun users or LOW risk selfies. The classification result is stamped on a document with an image of the visitors face and then sorted into one of two transparent boxes. The installation represents a security control situation familiar from e.g. airports. The recognition of threat is outsourced to assumable neutral machine.
Not machine P.O.V.