It’s the same problem across all decentralised protocols, if it’s cheap to say something you get spam(see email) but introducing costs can just skew it to those who can afford to spam instead (essentially those with an advertising budget).
So there’s been a lot of research on proof of personhood (BrightID/ideas/proofofhumanity) to add Sybil resistance mechanisms so we can do 1p1v across the network. They’re working ok, but the next big step is adding zkproofs so we can anonymise the voting (which is needed to prevent collusion) which clrfund and sismo are working on.
Kleros have an interesting curated register protocol, which seems to work on small scales. Some groups are using it to token rank guy issues to prioritise work and get feedback.
The status blog has some interesting writing around these ideas over the past few years https://status.im/research/