Is there some automatic way in Audacity to reduce damages done to music due to loudness war? Or in some other Windows software?
I do not expect it to be perfect- but to just ease / speed up process of further manual reversing of this loudness insanity, which is pseudo-dynamics and clipping. You know: to make a [typical] modern released recording sound more like it was released in the 1980s
I am using Audacity 2.4.2 [but if needed I can upgrade] on Windows 10 Enterprise 20H2 19042.746 x64
Loudness war was increasing amounts of compression.
The opposite of compression is expansion.
Here are 3 free expander plugins which work in Audacity on Windows …
La Petite excite … https://www.finecutbodies.com/?m=S008&p=sound [says “exciter”, but it’s a (2 band) expander ].
ToneBooster’s FIX … https://forum.audacityteam.org/t/free-v3-toneboosters-plugins/64538/2 (6 band)
Steve’s dynamic mirror … https://forum.audacityteam.org/t/envelope-follower-ducker/47090/6 (1 band)
Audacity has [u]Clip Fix[/u] which MIGHT help.
You MIGHT be able to improve the sound but you can’t really reverse the damage…
Sometimes compression is reversable if you know all of the settings/parameters. In the analog days DBX noise reduction used “compansion” and I believe it was also used in analog telephone systems. The signal is compressed before recording (or transmission) and expanded during playback (or receiving). That greatly improves the signal-to-noise ratio and dynamic range. It was never popular because unlike Dolby noise reduction it sounded terrible when played-back without decoding. And of course, digital made it obsolete.
But with music production there are several issues (besides the unknown settings/parameters)… With limiting (and clipping) it’s impossible to know the original height or shape of the waveform. Sometimes multiband compression is used (with more unknown parameters). Often compression is used on the individual tracks before mixing so that’s impossible to reverse if you only have the final mix.
…And, I suspect that much of it is in the performance, arrangement, and playing style. We’ve got a couple of generations of musicians & producers who think music is supposed to be constant-loudness. 
…We can’t have any quiet parts or any louder drum-hits or cymbal crashes, or anything that stands-out from the background!
Does anyone else has something to add?
Maybe there are some advancements in the field of audio AI and some new methods are being developed or recently have become available for usage?
Not yet… Maybe soon… Maybe in a couple of years AI will be able to remake the song from scratch with whatever changes you want!