Being in IT, I was already vigilant against spam and phishing. When I started my crypto journey and heard of all the hacks and dangers that come with it, I pretty much avoided clicking any links. I have been better at finding faults like the incorrect name, URL, etc. With AI, I do think these bad actors and their methods will only get harder from here.
It is going to get messy. People seem to only imagine all the "good" AI can do for them by making their lives easier, without considering what a bad actor will do with it.
I think with new technology, that will always be the case. If we can't imagine the good it can do, then why bother developing it. As with any tool, there is the possibility of it being used for bad things.