I have wondered, if all the data that the data collectors had could be used to create a confidence score for an individual (not just for ad revenue purposes). For instance, using studies, backgrounds, professional experience, interest areas etc.

If for example, I was giving a piece of information on coding, the algorithm could predict with some level of confidence whether I know what I am talking about. If you gave the same information on coding, the algorithm would favor yours, since you would have a higher rating than me, based on everything known about you. It doesn't make me wrong or right, but it gives you a louder voice in that area.

Given a different topic, it could be (big assumption here) that I get a higher rating based on my experience etc.

This would be like stake-based voting, except using experience as the stake, weighted based on the type of experience. Rather than one person, one vote, the most qualified in a given topic would have a higher weighting than the least.

Obviously, with perfect information, it would be possible, but even with imperfect information, would it yield a better overall result than one person, one vote?