3 ms·

So if it is, on one hand, "collaborative" and it expects people to work for free, and on the other hand there is apparently so much value in putting misinformat

• baxtr 4y ago
  Easy. We just need a collaborative fact checking system for the collaborative fact checking system.
• vntok 4y ago
  Paying some of those people to manipulate reviews wouldn't really work at scale. See here how Twitter automatically ranks reviews: https://twitter.github.io/birdwatch/diversity-of-perspectives/ https://twitter.github.io/birdwatch/diversity-of-perspective... > To find notes that are helpful to the broadest possible set of people, Birdwatch takes into account not only how many contributors rated a note as helpful or unhelpful, but also whether people who rated it seem to come from different perspectives. > Birdwatch assesses “different perspectives” entirely based on how people have rated notes in the past; Birdwatch does not ask about or use any other information to do this (e.g. demographics like location, gender, or political affiliation, or data from Twitter such as follows or Tweets). This is based on the intuition that Contributors who tend to rate the same notes similarly are likely to have more similar perspectives while contributors who rate notes differently are likely to have different perspectives.