On the contrary, I think the key is to consistently label nonconsensual taking of intimacies as "creepy" regardless of who does it.

I don't have a problem with that. My point is that some behaviors which aren't "nonconsensual taking[s] of intimacies" are still associated with creepiness (some for good reason, even, like murdering cats or what have you) and that we need to check those behaviors for alternative explanations (maybe they have really really bad allergies [although probably not!]).

Treating "creepy" as an XML tag on people is a terrible idea — both for the "creeps" and for others.

Taboo XML tag and say this again. Are you saying we shouldn't have a concept of "creepy people" only a concept of "creepy behavior"? Because while that might work well in the abstract, in reality it's important to predict how people will behave, which requires using labels and categories and heuristics.

Declaring, "Some people are noncreepy" is vulnerable to exploitation by people who manage to get themselves labeled "noncreepy" and then go around assaulting people with impunity, knowing that their noncreepy tag will protect them from accusations: when someone says "Hey, Joe assaulted me!" others will respond "He couldn't have done that; he's noncreepy!" That's part of what the feminists call "rape culture".

I think the position you argue against is roughly correct, actually. If there are behaviors and characteristics that strongly correlate with a lack of creepiness, it is logical to have a low prior probability on the judgement that someone with those characteristics will assault someone. For example, if someone is asexual or a child, then it's incredibly likely that they aren't a sexual predator. If someone is really nice and doesn't stare at girls or do anything weird, they're also probably not a predator. Yes, people can get around your best predictions, but that isn't a reason to not make predictions in the first place, it's just a reason to update when you get more evidence.

If Joe really isn't creepy at all then he's probably either innocent or a sociopath. If I knew Joe, and he seemed like a really nice guy, and he's never said anything that seemed slightly disturbing, and he'd been perfectly normal his whole life, I wouldn't just assume that he must be a rapist just because I heard a rumor somewhere. That would be dumb of me. The correct response is to gather more information, while for now going with your priors that say Joe is a-okay.

I don't like that you tried to stick me with defending rape culture, also. That's just inflammatory.

Not all actions which are labeled creepy have negative consequences or are exclusive to creepy people. If someone does an action without negative consequences but that action is then labeled creepy prematurely without considering the motive behind it, then this premature labeling will have negative consequences. Bad labels mean bad models of reality which result in bad decisions. Mistaking someone for a creep when they're not precludes making friends with a potentially good person, and means your predictions of their behavior will be wrong.

You say that "they are failing to take steps to change this even when the negative consequences are pointed out to them". I am confused. You seem to be envisioning a very specific type of scenario here, and I don't know why you think that I would be envisioning the same scenario. Since when did my argument get restricted to instances where someone has been confronted with their creepiness and refused to change? My argument is meant to encourage caution in the people here are considering experiences from their own lives and trying to determine who is and isn't creepy, on a general overall sort of level, not to this very specific kind of instance that you mention.