Either way, totally agree that it's a tragic situation.
The only sympathy I have for Google is that trying to separate the good vs. "evil" (as in "Don't be evil.") content is a monumental task that machine learning will probably never be capable of performing. So they're left with the choice of spending an inordinate amount on human review and detailed research or just make wildly over-broad removals.
I'd rather they leave up more rather than less, but they tried this approach and it almost lost them every major advertiser. So continuing down that road would potentially lead to the whole platform losing viability. Maybe some would like that outcome but these historical videos would be just as lost.
Maybe we'll see the pendulum swing back in an effort to reach a more reasonable middle ground.