That could work but at the end of the day it’s the guards who have physical control of the surveillance system. We would have to rely on this other system (powered by ML — who designed this system? - and independent auditors) working for prisoners instead of for the guards when anything goes wrong (sorry the recording malfunctioned during the exact period you claim you were assaulted).

Unfortunately prisoners are usually on the losing end of systems designed by their jailers, as they lack strong advocacy and a role in designing the system itself.

I thought this was supposed to be cheaper?