It’s often only obvious some new technologies are a fad in retrospect. It’s hard to tell how they will be adopted, Second Life had a very enviable early growth curve.

Perhaps, but only if you can get the correct buzzwords in. Such "research" is rarely about the impact of the tech itself (although there are always people trying to actually apply it, e.g. as a teaching or therapy tool), but more a way to express some overall theory (or set of theories) in a different context, as if it were an observation that can validate the theory. So yes, they jump on fashions.

It affects other subjects too. I know people who've done a PhD "in" Harry Potter or Sherlock fan-fiction, but it's not about Harry Potter or Sherlock. After all, there's not that much to say about it, but they both started from some "theoretical" point of view, and then just wrapped those ideas around whatever observations they can make up about the subject.

AI seems difficult in that respect. It would need actual understanding. Without that, you can't really make observations about it the way you can make them about some book or a game. Otherwise, it stays restricted to checking a few dialogs with ChatGPT and commenting on that.

Here's an example from someone who did a PhD on second life about ... the MetaVerse. https://www.linkedin.com/pulse/building-betterverse-note-div...

Just after finishing their crypto bender