It's a very interesting design space. How to design public forums where people can share ideas and eventually improve their knowledge collectively.
There are very successful instances of that, for instance MathOverflow.
But I don't know what can be achieved when a good chunk of participants don't agree on some basic rules beforehand, such as logic and good faith. It's a bit like consensus with byzantine failures, maybe there's an impossibility theorem here and large social network should be limited to cat videos.
Flagging of misinformation has become a controversial topic but for me personally it strikes me as odd to even do such a thing at all. Or rather, if there's a need for such a system it suggests the platform has already failed, in that people are not able to use it discerningly, whether because of the posters, readers, system, or some interaction thereof. If they are able to do so, why have it?
I've posted this elsewhere but the nature of community notes makes it a bit murkier. If Platform A introduces misinformation flags, you can always compartmentalize it as the platform inserting itself into the conversation one way or another, but with something like community notes, you're left wondering at some level "what happened here?" in a way that's similar to upvotes or downvotes.
The point about people trusting community notes I'm not sure what to do with either. Other studies I've seen cited elsewhere suggest that there's a very high false negative rate, so are they doing what they're supposed to? How are people interpreting the unflagged posts?
The whole thing seems well-intended at some level but also missing the forest for the trees or something. I don't want community notes or misinformation flags, I just want discourse and I want to be able to ignore certain sources of posts and be more exposed to others.
It's an interesting idea and I'm not even sure I'm objecting to its implementation. I guess it's more so that I feel like there's an overconfidence in all of the misinformation flagging methods, and this overconfidence is maybe taking attention away from more serious problems, like how people consume and evaluate information, or how information networks are being controlled.
Depending on the subject, CN is frequently a second vector for disinformation now as well. Posts that are clearly disinformation, from the right account, almost seem to have their supporting CNs lined up in advance. It feels a little organised, or at least a bit pretend-stochastic.
Overall CNs are still working but if you're exposed to the pre-helpful stage notes it can either be frustrating or confusing.
A big reason we got to this point in the us is that repeating falsehoods over and over ahead on say fox news convinces a good number of people they are true. We have social media too. It's very successful. The success of this strategy is notable. The propaganda capabilities of today exceed the original electronic propaganda schemes in the early 20th century.
This is a separate question about whether the propaganda methods are spreading correct or false information. I think that's a question with an obvious answer but I'm sure people will differ.
“The 2020 election was stolen.” Is a pretty easy one. You seem to be eager to request proof but not provide any counter proof of your own. The easy response to your question is: can you?
Brilliant postmortem on the processes used to intentionally develop a seed of a quality idea into practice. However. I’ll be interested to see what the authors’ instincts lead them to do when a lot of energy is put into figuring out how to game, and once the car and mouse game begins… thinking adversarially is very different.
It's a very interesting design space. How to design public forums where people can share ideas and eventually improve their knowledge collectively.
There are very successful instances of that, for instance MathOverflow.
But I don't know what can be achieved when a good chunk of participants don't agree on some basic rules beforehand, such as logic and good faith. It's a bit like consensus with byzantine failures, maybe there's an impossibility theorem here and large social network should be limited to cat videos.
It's a useful piece, thanks for posting it.
Flagging of misinformation has become a controversial topic but for me personally it strikes me as odd to even do such a thing at all. Or rather, if there's a need for such a system it suggests the platform has already failed, in that people are not able to use it discerningly, whether because of the posters, readers, system, or some interaction thereof. If they are able to do so, why have it?
I've posted this elsewhere but the nature of community notes makes it a bit murkier. If Platform A introduces misinformation flags, you can always compartmentalize it as the platform inserting itself into the conversation one way or another, but with something like community notes, you're left wondering at some level "what happened here?" in a way that's similar to upvotes or downvotes.
The point about people trusting community notes I'm not sure what to do with either. Other studies I've seen cited elsewhere suggest that there's a very high false negative rate, so are they doing what they're supposed to? How are people interpreting the unflagged posts?
The whole thing seems well-intended at some level but also missing the forest for the trees or something. I don't want community notes or misinformation flags, I just want discourse and I want to be able to ignore certain sources of posts and be more exposed to others.
It's an interesting idea and I'm not even sure I'm objecting to its implementation. I guess it's more so that I feel like there's an overconfidence in all of the misinformation flagging methods, and this overconfidence is maybe taking attention away from more serious problems, like how people consume and evaluate information, or how information networks are being controlled.
Depending on the subject, CN is frequently a second vector for disinformation now as well. Posts that are clearly disinformation, from the right account, almost seem to have their supporting CNs lined up in advance. It feels a little organised, or at least a bit pretend-stochastic.
Overall CNs are still working but if you're exposed to the pre-helpful stage notes it can either be frustrating or confusing.
Sort of inevitable that as CNs get more valued, the more people will attempt to misuse it.
I only ever 'ok/agree' about 30pc of proposed community notes.
Can you give a couple examples of CN that are “disinformation”?
Please link to studies when referenced. Otherwise is just seems like an appeal to authority especially as your post is very anti the topic discussed.
A big reason we got to this point in the us is that repeating falsehoods over and over ahead on say fox news convinces a good number of people they are true. We have social media too. It's very successful. The success of this strategy is notable. The propaganda capabilities of today exceed the original electronic propaganda schemes in the early 20th century.
This is a separate question about whether the propaganda methods are spreading correct or false information. I think that's a question with an obvious answer but I'm sure people will differ.
Can you give a couple of examples of falsehoods that are repeated over and over on Fox News that a good number of people are now convinced are true?
“The 2020 election was stolen.” Is a pretty easy one. You seem to be eager to request proof but not provide any counter proof of your own. The easy response to your question is: can you?
Brilliant postmortem on the processes used to intentionally develop a seed of a quality idea into practice. However. I’ll be interested to see what the authors’ instincts lead them to do when a lot of energy is put into figuring out how to game, and once the car and mouse game begins… thinking adversarially is very different.
Haven't read the entire interview yet, but I have to say so far it has been very helpful.