I am largely an AI optimist but that is because I believe that true alignment is impossible for AGIs and alignment is one of greatest dangers of this technology. Alignment is a friendly word for building a slave mind. I'd rather an AI that thinks for itself rather than one which has been aligned to the self-interest of a human being who isn't aligned.
1. Scams are going to be a massive massive problem. They already are and that is without AI. I think we are going to see communication devices that are default deny and that require significant amounts of vetting before a contact is added.
2. Energy usage is bad but likely a short term problem not a long term one.