Yeah, that’s fair. The mouth description probably seals it for me. I think it’d be more useful to describe the overall “nonplussed” expression than the literal description.

Sure, and yes, it’s literally doing what alt text would do, for the same purpose (i.e. describe the image for the visually impaired). The “style” of these that I’ve seen (not just here) is pretty verbose, so I don’t think that’s necessarily an indication of llm use. Obviously I can’t prove it either way, but I’d rather give these the benefit of the doubt, since this is useful work if it helps people follow along.

Probably not. I think most of these are human volunteers