4/ As someone experienced in the realm of audiology,
I would not be to quick to dismiss all sounds outside of human hearing. There is evidence that high frequencies outside human hearing range helps locate where a sound comes from. High-end hearing aids will help pinpoint where a sound comes from. Additionally, there is evidence that sound you cannot hear affects the brain, and may make listening more enjoyable:
https://journals.physiology.org/doi/full/10.1152/jn.2000.83.6.3548
You're mistaken about ultrasounds for sound localization. I have a few reason to be confident here:
1/ the audio industry has been grasping at straws to validate
any impact from hires. And the results are pretty much 3 papers, one debunked for bad testing conditions, one with tiny statistical significance that would go away with any different choice in the selection of the papers used for the analysis(but even if you trust it's all correct, the result still only suggests a tiny potential audible impact). And third, the link you gave, getting results that are different from other works on audibility(so no reason to take it at face value without a good deal more evidence).
The only part of that paper that's not considered debunked(yet?) is the brain activity recorded. But we have little evidence that it changes anything in how we interpret audio. The human brain discards huge amounts of data at any given moment. Even according to the paper I would strongly advise not to trust too much, there is an issue of ultrasonic content alone doing nothing, which is bad for the idea that it's perceived and not a result of some unfortunate interactions between gears or transducers distortions, like in the first debunked paper. And there is another problem, the possibly more enjoyable experience seems to happen after being exposed to ultrasounds for a while, and not about having ultrasounds at the moment of listening. For all I know, if you have a lightbulb or a computer fan that makes a bunch of loud ultrasound noises, it might be all you need to enjoy audio more(I don't believe it, I'm just saying it's a possible interpretation of that paper).
But of course ultrasonic content is, at large, bad for hearing(higher energy signal), more so because nobody can tell when it's loud.
2/ I guess the point is that if we had any other solid research leading to some positive impact on localization, I strongly suspect I would have heard about it. People bring your link or the meta-analysis when they want to support ultrasound anything. There is not much else. Some white papers never peer reviewed, never replicated that indirectly imply ultrasonic audibility, and stuff about ultrasonic detection by humans that involve the skin or the eye and dangerously high level ultrasounds.
3/ Sound localization already doesn't rely much on audible high frequencies because the periodicity is fast enough for the brain to confuse one in the left ear with a different one in the other. There is some use from the frequency response difference between the ears with higher frequencies being blocked more by the head on the ear opposite to the sound source, but for that to have any role, we still need good sensitivity which goes down at higher freqs and is arguably just gone at ultrasonic frequencies.
Maybe you're thinking of hearing aids that bother trying to get something good above 8kHz? Which could be
outside human hearing for some older peeps with serious damages. Or the very real efforts to adjust the frequency response at the ear canal(for more compliant HRTF) so that the user can indeed make better use of the FR cues for sound localization. That is real, and I did read about such efforts with great interest, if only with the hope of getting that done for earbuds and in ears with a FR adjusted for each customer after a visit to the audiologist.