The question is, how are such AI devices to be 'trained'? As I understand it, they use 'learning algorithms' which analyse many examples of whatever they're trying to replicate.
To get suitable examples, one would need to attach sensors to a large number of spankees, and their spankers, and take careful measurements of the effects (both physical and behavioural) of spankings of varying intensities, with a variety of implements, on a large selection of spankees.
For calibration purposes, a statistical cross-section of spankees would need to be subjected to a series of spankings to test the effects of different intensities of spanking on the same bottom, while the effects of implements would need to be tried on a range of spankees' bottoms, in order to find the most effective.
Given the difficulties and cost of this research, I think that it will be some time before skilled spankers will be superseded by AI robots! | 
Page 2 of 2: «« 1 