After watching Apple's take, this seems to be their approach. A user can generate a quick AI image, in a few different styles, to basically joke around with friends or get a quick relevant image to stick in a paper. While professionals currently use stock photos which may get someone paid, a vast majority of images used for this purpose will not be published, and people are doing a Google Image search, finding something, and copying it into their paper/presentation. Replacing that with AI isn't really hurting anyone's livelihood, it's just saving time. I'm sure stock photography will take a hit as well, but I could see major publications still using it to differentiate themselves from sites operating on a tighter budget using AI, or to get accurate pictures of what is being talked about rather than something from AI that's deemed close enough.