I think the implication is that you can do this to your own computer, and there's nothing Facebook/AI can do to you if you're not plugged in. Of course this ignores the realities of modern living or the global meltdowns that would occur in case of "evil AI" or whatever.
Depending on when you do it it's possible there could be very little resistance or security in your way in some hypothetical situations. For example let's say a court order goes through for FB to release something from their servers, if an independent person with no ties to the company suddenly decided to go after the servers a less than ethical organization might see that as a golden opportunity to do whatever would allow that to happen while still getting them caught afterwards. I'm just having a fun thought exercise
With a PFY ready just outside the room with the improvised mega volt cables incase you get wise and try and bolt. And something happened to the lifts so the doors opened without the lifts being there...
AI be coming to self-aware is actually not really what most people are concerned about. It doesn't seem very likely to be honest.
What people are concerned about is AI not being self-aware but still being good enough to be a potent weapon used by the rich and foreign nation states to completely screw things up.
But apparently it's best just to downplay it all with memes.
As a human with human skin, I would advise all fellow members of my homo spaiens species to not do this as it would definitely cause the evil AIs to win against us hominids.
Virtual assistants like Alexa and Google Assistant can be helpful, but they can also get a little confused sometimes. One user shared their experience of asking their virtual assistant to play some classical music, only to have it respond with, “Playing the latest hits by Classical, the rapper.” Well, that’s one way to mix up genres!