[root_axis]

> The surveillance AI knows nothing of the difference between malum in se (crime!) and malum prohibitum ("crime")

And for the humans in charge of the AI there is no such distinction so the AI is working as intended.