If the active layer is a camera layer, the view from that camera will be shown. Otherwise, the view uses the bottommost camera. So you can change which one is active in the view by rearranging the layers. When exporting, you have the option to choose which camera to use for the export.
Multiple cameras could be used to create different perspectives/camera movements for the same shot. It could also be used to export multiple different regions from an animation, for example exporting individual sprites from a spritesheet. Duplicating the camera layer can be used as a “backup” before testing changes to the camera. There’s probably other uses, but ultimately it’s up to the user to decide how they want to use the software.