These kinds of rules are never applied strictly. Content moderation is always discretionary. If you don't want to be judged as adult content then play it safe, which includes censoring the f-bomb. It's just common sense.
This is more than a one dimensional “just don’t have any curse words or non family friendly content in your videos because it’s entirely your own fault if YouTube flags it capriciously ” kind of issue.
Too many times I’ve heard of this failure mode. Did humanity fundamentally lost against AI?
They're probably running scripts on the content and likely not basing decisions on human review of any kind.
If they have any human review, it's probably contactors, reviewing only the 3 second clip of the 1.5 hr video, the tiny part that was flagged by the algorithm and understanding very little.
There is not even a word spoken in it.