of course, there are those who believe in the efficacy of "upsampling" ...
I wonder if that is correlated with being an anti-vaxer?
OTOH, there are some AI based fill in the dots products for post-processing in photography coming out - I think Adobe has one. One could certainly do that in audio - and maybe has already been done for tics, pops, etc.
But, in general, the best study on this is the meta-analysis conducted by Reiss. While imperfect, he found only a marginal improvement.
A study using very high res vs. Redbook or even mp3 with young musicians might be illuminating.
I suggest sticking to speakers/room, the best masterings, and finally maybe power amps for better SQ.