[troutman]

Originally Posted by Fuzz

Go read the WWW trilogy by Canadian Robert Sawyer, and get back to me on that...

https://www.quora.com/How-do-I-defea...-were-a-threat

AI is bound to several rules:

• AI is computer-run, meaning that a digital scorched earth campaign would destroy it, or confine it.
• AI is programmed, meaning that its algorithms and processes are all derived ultimately from one program.
• AI cannot function without power, meaning that a shutoff to its main system could be lethal, or at least very damaging

https://en.wikipedia.org/wiki/AI_control_problem

AI risk skeptics believe that superintelligence poses little or no risk of accidental misbehavior. Such skeptics often believe that controlling a superintelligent AI will be trivial.