NetHui 2020: Arbiter: A New Disinformation Fighting Non-Profit (Jon Bell)

Oct 13, 2020 05:19 · 685 words · 4 minute read voting model sift first product

  • [Jon] Hi. My name is Jon Bell, and last month I started a non-profit. It’s designed around fighting disinformation. Our first product is going to be a newsreader and that’s what I’d like to talk to you about today because I need help. I need people to talk to about this and I need people to beta test this product to make sure it’s as good as it can be. But, hold on. A news reader? How in the world does that fight this information? Well, I used to work at Twitter, fighting disinformation, and we made a bunch of features.

00:30 - And the features that we worked on all basically had to do with a pretty simple concept of take the default Twitter experience and filter it. Remove this person, downvote that tweet, et cetera. It was all around filtering and downvoting. The issue, however, and the reason that it doesn’t seem like Twitter has made a whole bunch of progress, is I think three things that all social media companies are dealing with. Number one, they are a for-profit company so they’re always going to be putting that ad revenue before anything else because they need to keep the lights on.

00:57 - Number two, it’s all based on algorithms that are trying to figure out how to get you to click ads rather than making sure you’re as informed as possible. And number three, it’s a popularity contest. On any one of these sites from LinkedIn, to TikTok, to Reddit, to Instagram, every one of these sites, there’s a follower and a voting model, and that opens up all sorts of vectors for abuse and disinformation and manipulation. So we thought, “All right, what if you flip it around?” What if it’s a non-profit, so everything we’re doing is for the benefit of you? More like a Wikipedia than a Facebook. And number two, instead of algorithms what if you could customize that algorithm and see exactly what’s happening so there’s no black box at all? And then finally, instead of having a follower and voting model, what if that’s just done away with and you’re just left with a great newsreader? Imagine being able to go to Facebook and say, “Hey, Facebook, on your timeline I don’t want to see so much anger or politics.

” 01:49 - Imagine being able to go to Reddit and say, “Hey, I just don’t want to see clickbait anymore. If you know it’s clickbait, just don’t show it to me.” What if you could go to Twitter and say, “Can you please optimize for optimism and New Zealand news?” And let’s say it would go off and do that. Well, it doesn’t do those things today and so here’s our idea. If you’ve ever used RSS, step one would just basically be like RSS. Bring in various news sources and then it just aggregates it into a list of links. Very simple. Then, you go into the system and you just say, “Hey, I want to see less of this, more of that, and I want you to block these things outright.” Then, the following thing happens: That list that you start with starts sort of upvoting certain things that you say you want more of, blocking things that you said you didn’t want to see at all, and then downranking things that you said you wanted to see less. The end result is that this thing on the left is what we’re used to seeing today, where we have to sift through everything. The thing on the right is a customized, personalized, transparent filter so you can see more of the things that you want to see.

02:48 - So that’s what we’re working on, and that’s where you come in. If you think this is interesting, if you’d like to chat with me about this, this is the point in our development where we would love to hear more peoples’ ideas and also get some beta testers. Thanks very much! If you’d like to learn more, please go to arbiter.space. .