IMO there iseems to be a lot of misconception and generalizations on AI. Traditional software is a set of instructions that are deterministic where (the same input always yields same output). Whereas AI software instructions are probabilistic (using accumulated data in decision making) and can improve the decision making process over time as they are exposed to more data. The old adage "garbage in, garbage out" is even more valid with AI.
Seems every software program out there now is touting AI functions, but I have to wonder just how much learning capability they actually have and how much is marketing hype. AI editing tools are very specific function apps...a denoise AI tool won't compile a text to image, but it will use compiled data to sharpen details. Furthermore they all require human input on some level. Much like a mentalist AI makes guesses from tells, the more tells, the better the guesses.
There's alot of hand wringing, gloom and doom, but once the newness wears off, I think you'll find as Sharon indicated people will still do their own thing.