This isn't useful or sufficient. You have to consider how many bots get banned and cost to determine efficacy. If you want 10,000 fake people to manipulate real people $10,000 doesn't seem a high price if you make the fake people act organic enough that they largely aren't banned.

It would be more useful if a singular service verified sufficient credentials to prove you were an authentic human and allowed you to auth to various sites. This in turn creates the problem that verifiers now know a LOT about your online life.

If the verification involved site -> verifier -> government held public key I think you could arrange so that none of the parties had enough info to identify users.