Have you tried to make a throwaway Google account lately? It's kind of useless to do so. Not so much in the sign-up, but due to the fact that accounts with no history and no clear ties to a physical identity have nothing "vouching" for them, so constantly trip the "Are you a robot?" checks. (And, I imagine, {email, comments, Google Docs shares, etc.} from such accounts also are invisibly thresholded lower on any spam filters they are run through. This is why people write viruses to take over people's existing Google accounts—accounts without "reputation" are kind of worthless for doing anything public-visible.)
This type of system solves a lot of the throwaway-for-purposes-of-evil-behavior problem, and may well be able to do that algorithmically, in lieu of human moderators. In return, it requires people to give up a lot of privacy/anonymity to participate. I don't think that's an inherently good or bad tradeoff, but it is definitely one people should be aware of.
> ensure comments meet a basic level of decency
Decency is famously difficult to define [0]. One person's idea of decency is another's idea of "free and open debate", is another's idea of alternative culture, etc. And the truly bad actors are using throwaway accounts anyway, or will start as soon as you start banning them.
I think active moderation and community curation is the only real solution to this. Sub-communities (formed around particular channels or topics) have to define and enforce their own standards of decency. In this case, I can understand Youtube taking the conservative approach of not trusting communities to do that, especially given a lot of the stories that have come out showing that they aren't.
[0] https://en.wikipedia.org/wiki/I_know_it_when_I_see_it