YouTube isn't consistent in how it applies warnings or strikes. Going through both of my channels, one of which contains mostly air check recordings which contain some copyrighted music, there's no consistency.
In some cases YouTube apparently has an agreement with the music license holder to allow videos to be published, with YouTube inserting ads. Presumably any revenue goes toward compensating the music owner. That's fine.
One of my old attempts at creative stop-motion photography used "Little Miss Echo" by Raymond Scott as music. At the time I made that video YouTube's online editing tools were terrible and there was no selection of license-free music. So I chose some obscure stuff that might escape their AI. It did, for awhile. It still has a copyright claim, but it's still viewable and the music hasn't been muted. YT inserts ads to compensate the owner. Fine with me.
Another is a demo of the Palstar SP30 speaker vs a Radio Shack Optimus XTS3 speaker, using whatever music was playing on some shortwave station that night (probably WTWW, IIRC). That video got a copyright claims for "Music! Music! Music" by Del Wood. But the copyright owner had an agreement with YT to allow it. Presumably YT enables ads on my video. But I can't monetize that video. No problem, I don't monetize any radio related videos for that very reason. Demos inherently involve the possibility of copyright claims.
An antenna demo video got two copyright claims, one for a Romeo Void song (hey, I love Debora Iyall, so if it benefits her in any way, go for it). But the video wasn't blocked or muted.
But few of my dozens of other videos and audio recordings of jukebox pirates like Wolverine, Twentieth Century Radio, etc., has ever been hit with a copyright claim, despite entire songs in decent fidelity being recorded. Who knows.
Another is a clip from an old Motown music awards program, recorded with my video camera of Mary Wells just after she died. Mary was wearing this magnificent puffy blue gown that took up practically the entire stage. I got the impression she was making a statement after having been disrespected by Motown for years. And they cut her performance to a medley less than 60 seconds long. I couldn't find a video clip anywhere of that particular brief performance, which was a shame. So I uploaded it, fully expecting a warning or copyright strike. Years later, it's still there.
But a short clip from a nighttime group bike ride with friends, with Cake's "Long Skirt, Short Jacket" playing in the background on cyclist's Bluetooth speaker, barely audible over traffic noise, got a strike.
I'm guessing the warnings, strikes, audio muting or deletions are applied to videos containing music for which the artists or license holders haven't agreed to any such uses.
After that I started using my own compositions, mashups and remixes as background music for my videos. Much of it is copyrighted material, but it's unrecognizable to the current level of AI (YouTube, Shazam, Google, etc.) because of back-masking with overlapping sounds from various sources, some original material, and lots of effects via Cool Edit Pro or Audacity. So far, so good, no warnings. I've noticed that Shazam and Google music recognition AI struggles with shortwave pirate station music unless the signal is nearly perfect. Any static or hiss and often Shazam misses. Google works more often on recognizing songs despite static, but still fails about half the time.
But judging from complaints by many YouTubers who create original content for a living, there's no consistency inYouTube policies. It seems as if YT is using AI to scan every upload and if it perceives any controversial content (firearms, discussions of firearms laws, some political or medical/health issues, reasonable editorials on hotbutton topics or even parodies), the bots pre-emptively block the content and issue strikes. It's up to the channel owner to defend their content, and hope the appeal actually reaches a human rep at YT. I follow all kinds of channels from every political and cultural extreme, and they all get random, unexplained warnings, strikes, etc., so there's no apparent bias by YT. Just bad AI attempting to replace informed human moderators.