I’m not an expert but how I see is it is that behind every AI there is a code and behind every code there is person and behind every code or AI that write a code there is person who write it behind it.

I think there can be a “malevolent” AI but it’s not going to be like that because of itself, but because someone programmed it to be like that. If you code something big, with a lot of variables to the point that it look like is thinking by itself BUT is not.