I wouldn't get a chip simply for the fear of malfunction whether it be unintentional (a software bug) or intentional (hacking).

I also don't really see the need to integrate with AI either. AI is simply mathematical processes running with data. It's fundamentally stupid like your computer is, but it "appears" to be intelligent.

If you don't want your AI to misbehave, then program it so it won't misbehave. An AI is not going to emerge out of nothingness, someone is going to have to program to do bad things or learn how to do bad things. In any case that is the programmer's fault and not the AIs. I just do not think we have to worry about AI. We have to worry about other people.

Good article, btw.