What level of quantitative evidence would you like to see?
Would you prefer if replaced "no longer useful" with "has diminished significance such that it is ultimately vestigial"?
In my world it is plainly obvious that music discovery occurs from algorithmic introductions, which are counted, and the song being played which is also captured by audio listening apps which is also counted.
In the past, cover art would have been an advertisement for an album that contained songs which weren't out yet, and also useful when browsing physical albums in a store. Both of which is pretty much not happening at all, I mean feel free to correct me if your world still has that. I remember people used to complain about all the other songs on albums not being as interesting as the one song they wanted, the people are the market and the market chose something else which evolved to a completely different form of music discovery.
I don't have a paper on that. Is it really moot if this is a shared experience? I think there are a lot of people that are just uncomfortable that the world changed and they never stopped to notice, and without a counterpoint thats how you sound to me. But if that's not the case, I'm totally open to a conversation as I don't have strong opinions on the matter and am also totally content with music discovery today, which doesn't emphasize cover art for easily understandable reasons, to me.