The speaker argues that a human should watch the footage from a dashcam and determine whether or not a fine should be issued instead of a machine algorithm. The speaker provides an example of a situation in which the context should be taken into account when reviewing footage.