Reminds me of the Overwatch moderation system used by Valve to collect cheater data from human evaluators in Counter-Strike. Eventually, they leveraged that data to improve their automated cheat detection systems. [0]
This was actually a genius system on Valve's part. I think what makes these two systems different are a few things:
-Overwatch at least had some method of selecting people that "knew about the game." People that had never ranked competitively in CSGO couldn't make decisions if someone was cheating, whereas it seems that the hopes for this new Twitter feature will be that anyone can "fact-check" a tweet. Even if evidence is required, it is beyond easy to find secondary sources that skew facts or statistics into a different connotation.
-Whether or not someone is cheating is pretty binary. They either are or they aren't. One of my biggest concerns for Birdwatch is that it was likely be used on tweets that aren't binary statements of "write/wrong" facts, and will likely be used in mass-reports of those of other political stances. That, or shitposting
Like you're going to hold the Leninists accountable for their unanswered claims against the Breadtubers for today's Voash drama?
Or intervene with a fact check when someone says the 8052 is UV reprogrammable when only the modern clones are?
Or for when the guy who talks about Magic the Gathering and was in Mulan makes an offhand comment about the new Harry Potter series and literally everyone gets upset?
The point here seems not to have independent professional fact-checkers, but to crowd-source it.
On the one hand, I certainly struggle to see how more crowds will improve upon what is fundamentally a problem with crowds to begin with.
On the other, Wikipedia seems to work reasonably well--and purely as a result of the community mores, not due to any sophisticated moderation algorithm or ranking structure or similar.
I'm not holding my breath, but it is interesting to see if a community-driven effort imbued with a Wikipedian-like spirit might succeed where more automated efforts have not.
The underlying question is philosophical freedom vs authoritarianism. The same tactics can be utilized by different actors who have different worldviews. The Stasi utilized Crowdsourcing to enforce authoritarianism
for example. One person's terrorist is another's freedom fighter. One's misinformation is another's truth particularly when it comes to how information is interpreted & what is focused upon.
Once a proposition is falsified a rational community would rejected it.
Once your community or protocol "philosophically" rejects checksums it is only a matter of time before your system breaks catastrophically.
This philosophical freedom vs authoritarianism thing you're proposing is a view that is incomplete and small to the point of being false.
The actual underlying question is do you accept that grounded truth exists and are you willing to have a large and nuanced enough view to include all non false propositions and reject all false propositions.
Sure, and if you squint hard enough, there are similarities between having a warning label slapped on your Tweets and being thrown in a gulag for decades.
But if you're squinting that hard, not a lot of light is getting through.
> On the one hand, I certainly struggle to see how more crowds will improve upon what is fundamentally a problem with crowds to begin with.
When you only have a hammer, every problem looks like a nail.
I believe that the social networks have built technologies that probably can't actually be made to work in a socially beneficial way, since they're utterly dependent on crowdsourcing and automation for economic reasons, but those tools aren't actually up the the problems they face.
Wikipedia works because you can remove content (even remove accounts afaik) once you're sure it's providing wrong information. On Twitter, people will interpret this as censoring.
Personally I have found Wikipedias quality to be declining for anything with a political slant to it. Articles about any public figure that has a right wing opinion looks more like a smear piece than an objective piece. Often phrases like "has been described as" are used in place of objective facts.
The problem with this statement is that it assumes that "fact-checkers" are a single entity when that is not the case.
The answer to "who fact-checks the fact-checkers" is "the fact checkers".
And we can use consensus, reputation, and checks against known facts to validate the checks themselves. If group A, group B, and group C all agree and group D disagrees, then we can say with relative confidence that D is probably wrong in this case. It is safe to go with the consensus.
However, if A, B, C are notoriously unreliable and have a history of distorting claims, then it is beneficial to see why D disagrees with all three. Or even not use A, B, and C as sources.
We can also judge a fact-checker's veracity by how they evaluate a claim. Do they provide sources for their claim. It's one thing to say "This will murder 40 people a minute" and quite another to say "This claim disagrees with a recent study by the CDC [link to study] that says it will only murder 39 people a minute".
The only reason to cast doubt on the process or activity of fact-checking is if you don't want people checking your facts.
Would be very interested to see how this addresses one of the primary underlying pathways of misinformation - confirmation bias. Many people believe information because it confirms their worldview; whether it's provably true or false is often irrelevant. In fact, I suspect that having a belief proved wrong often even reinforces that belief in some cases.
How much does truth matter in a post-truth society?
Truth still exists whether it’s completely hidden. Truth still exists if bad actors have injected 10 other believable stories alongside it. Truth still exists even if every fact checker calls the claim false.
People have spent enormous sums of money and time to hide the truth from the public. Just because every effort has been made to hide the truth, doesn’t diminish the fact that the truth matters. In my opinion, it’s just the opposite. It matters more than ever.
> Truth still exists whether it’s completely hidden
That kind of depends if you are discussing something objective or not. Wikipedia seems to have developed a strong left wing bias on anything political these days. Articles without any kind of politics still seem to be pretty objective. Can there be a single truth in politics, where ideas are very subjective?
Whether you believe in Covid or not if you inhale the virus you may develop the decease. If you don't believe in guns you may find some nasty people in your house who do, and you will not be able to defend yourself.
The truth still reigns supreme, it is just that people sometimes suffer unnecessarily because they do not know what the truth is.
It's much worse than that--those people who are, shall we say, misaligned with the truth impose consequences for others. Especially when those people are in prominent positions in government and culture.
I'm curious what mechanisms will be in place, and how effective they will be, at preventing dogpiling on people using this system.
As a specific example, as a nonbinary person, I'm constantly running into people online who tell me there are only two genders, or that singular they is some sort of new concept. My concern with a consensus system is that it will be used to shut down people like me (or trans folk, or socialists, etc. etc. etc.)
How do you build a consensus system that protects minority persons and also weeds out misinformation?
I have never come across a gendered language with more than two genders. If you are referring to "gender identity", it would be a good idea to use that phrase rather than the ambiguity of "gender" which now means different things to different people.
Gender identity is a different concept. It's an internal conception of what gender you are. Compare that to, for instance, gender presentation, which is an outward facing set of traits that correspond to which gender you are showing others.
Gender, in the context of language referring to a person, has a fairly settled meaning.
As someone classed as middle aged, gender is a synonym for sex to me, as it had been for decades. If you want to introduce new concepts, it would be easier for everyone if you used new words instead of trying to take over existing words.
This is the central question. If someone isn't asking, "What would a malicious user do with this system?" and inserting their favorite internet bogeymen as a thought experiment, they haven't finished their design.
This is doubly true for safety systems. I don't see much of the press release focused on how abuse of the safety system will be avoided beyond, "We'll eventually let it be reviewed by outside parties."
I'm pretty sure this project will just turn into opposing groups marking anything they disagree with as disinformation, the bigger group succeeding in having control of the narrative.
What's to stop a flood of Russian accounts gaining control of what's marked as disinformation?
I liked some blockchain / crypto based ideas people were having - like you have to use crypto to even say something is fake or not. A journalist has to "pledge" crypto that a news is not fake and the smart contract releases the amount back based on certain evaluations like if the post was voted as fake or not etc. (I am sure as described there will be holes, but I like the direction it was heading.)
> As we develop algorithms that power Birdwatch — such as reputation and consensus systems
Consensus is the enemy of understanding. For topics where there is conflicting or poor evidence or that bear on the culture war, I do not want people voting on what is the consensus truth. I want to see all the evidence. We have enough problems with researchers not publishing uncomfortable data; I don't want the little that exists flagged because it conflicts with the average Twitter user's sensibilities.
> Consensus is the enemy of understanding. For topics where there is conflicting or poor evidence or that bear on the culture war, I do not want people voting on what is the consensus truth. I want to see all the evidence.
That's a nice sentiment in a vacuum, but we're not in a vacuum. It's sort of like focusing on getting every child a college education when you can't even manage to teach them all literacy yet.
At the point, I feel like the priority needs to be around figuring out a way to defeat disinformation and misinformation to restore some kind of common ground understanding of the facts. The marketplace of ideas can't function if major factions reject truth-seeking in favor of blatant lies that are appealing for various other reasons (e.g. emotional satisfaction, ability to manipulate others, etc.).
We can all agree that chocolate is the best form of ice cream. Those who would disagree are obviously trolls spreading disinformation. The marketplace of ice cream can't function if we have other flavors like pistachio or strawberry being labeled "best" instead of the only truely best ice cream. Obviously these strawberry promoters are just manipulating other gullible ice cream consumers for their own ends.
> We can all agree that chocolate is the best form of ice cream. Those who would disagree are obviously trolls spreading disinformation. The marketplace of ice cream can't function if we have other flavors like pistachio or strawberry being labeled "best" instead of the only truely best ice cream. Obviously these strawberry promoters are just manipulating other gullible ice cream consumers for their own ends.
You're confusing preferences and facts. I was talking about facts [1], you're talking about preferences.
[1] like Donald Trump is not waging a secret war to defeat a conspiracy of Satan-worshiping pedophiles who harvest a (fictitious) drug from trafficked children and Trump lost the 2020 election fairly.
> The marketplace of ideas can't function if major factions reject truth-seeking in favor of blatant lies that are appealing for various other reasons (e.g. emotional satisfaction, ability to manipulate others, etc.).
The problem always becomes that few think their faction has this problem. Most people think that they and their side believe in the truth, and that the other side is peddling false information. In such an environment, "combating misinformation" inevitably becomes "trying to get everyone to agree with me."
In order to combat misinformation we'd probably have to start by combating the emotional commitment we hold to a certain narrative before we even examine all of the facts. That's not easy to do, but I imagine a good first step would be to cut out the constant 24/7 news cycle that a lot of people are addicted to and that seems to feed into these emotions. Twitter, of course, is a large source of this stuff.
Getting away from the news is an excellent way to start thinking more clearly.
> The problem always becomes that few think their faction has this problem. Most people think that they and their side believe in the truth, and that the other side is peddling false information. In such an environment, "combating misinformation" inevitably becomes "trying to get everyone to agree with me."
The problem with your line of thinking is that it assumes good faith and a certain amount of competence, which are assumptions that I don't think we can reasonably make anymore in light of the insane success of things like QAnon and "Stop the Steal."
Maybe I am interpreting it wrong, but this seems like it provides what you want. It lets people provide whatever evidence is relevant in response. People provide responses to the tweet, with the evidence in the response. And then are rated on if they are helpful or not. Its certainly not perfect, and probably going to be subject to brigading if people don't like the evidence provided though.
I suspect this will be minorly used for direct misinformation control("Donald Rumsfeld is not a lizard person"), but mostly used for narrative control("While this fact is true by itself, you need to look at the bigger context..."), and so will be entirely ineffective. People readily take up new information, but are very hesitant to change their internal narrative on a matter. Especially when they're being told by others what their narrative should be. Double especially if part of their narrative is that big tech/liberals/academia/coastal elites/etc are trying to feed you the Big Lie.
Conservative misinformation is a big talking point for liberals, and maybe the big societal issue at the moment, but as a small scale test run, I'd love to see birdwatch try to correct the record for misinformation that is commonly believed in liberal circles: Anti-GMO, anti-vaxx, toxic whiteness, the extents of systemic racism in our society, some of the more dire prognostications of nuclear war and global warming.
This will be a kangaroo court used to blunt-force-trauma a select subset of strawmen and deplorables off of the service. In the end, does anyone really care if Twitter is "truthful"?
> In the end, does anyone really care if Twitter is "truthful"?
As it's becoming an ever-increasing method of communication between government officials and whatnot, I'd say some amount of "truthfulness" and "validity" is warranted. The other day a member of the US government (I think it was Ted Cruz? I can't find it at the moment) wrote a tweet that implied he thought the Paris Climate agreement only affected the citizens of Paris. Something as factually "write/wrong" as that should be able to be contested and held accountable to being "wrong"
That being said, I agree with you that this will not be used as intended and will undoubtedly lead to more confusion and confirmation bias from all sides. Using crowd-sourced knowledge in this manner leads to the assumption that "what the majority thinks is right is factual," and I'm reminded it was once "fact" that the sun rotated around the Earth
On the surface, it sounds really nice. But anyone who’s paying attention in the fake news era has a general idea on what this will be used for.
Any tweets questioning the official narrative will be roundly criticized and ridiculed. Instead of having to delve into a long reply thread to see debunkers make their case, there will be an easy to digest notice within grasp. Which is not necessarily a bad thing if it is used fairly and responsibly. But judging on Twitter’s past performance, it likely won’t be.
Dissent will be publicly humiliated while pure disinformation from governments, think tanks, and corporations have no such objections. Any doubts to the accepted story will be pointed at one of the fact checker sites and further inquiries censored.
In an honest world, this would be one giant step to finally getting at the truth. But this isn’t an honest world, is it?
> In an honest world, this would be one giant step to finally getting at the truth. But this isn’t an honest world, is it?
Not to mention that in an honest wold, why would you even need something like this?
You don't need it in an honest world, and it will end up being worse than nothing in a dishonest world. It doesn't seem like a good idea in any scenario.
> Not to mention that in an honest world, why would you even need something like this?
The world is what we make it - we are the masters of our own destiny. This is literally true (subject to boundaries imposed on us by nature: the laws of physics, the physical resources available to us, the nature of the evolved human mind and the societies we have built for ourselves to operate in, and other things I may overlook).
If a variable in the world is not to our liking, and it is fixable, we can choose to fix it (as a collective species), or not. Mother Nature imposes some restrictions on this, but not many.
From the article:
> To date, we have conducted more than 100 qualitative interviews with individuals across the political spectrum who use Twitter, and we received broad general support for Birdwatch. In particular, people valued notes being in the community’s voice (rather than that of Twitter or a central authority) and appreciated that notes provided useful context to help them better understand and evaluate a Tweet (rather than focusing on labeling content as “true” or “false”). Our goal(!) is to build Birdwatch in the open, and have it shaped(!) by the Twitter community.
> To that end, we’re also taking significant steps to make Birdwatch transparent:
- All(!???) data contributed to Birdwatch will be publicly available and downloadable in TSV files
- As we develop algorithms that power Birdwatch — such as reputation and consensus systems — we aim to(!) publish that code(!) publicly in the Birdwatch Guide. The initial ranking system for Birdwatch is already available here.
> We hope this will enable experts, researchers, and the public to analyze or audit Birdwatch, identifying opportunities or flaws that can help us(!) more quickly build an effective(!) community-driven(!) solution.
If what they are saying is 100% true (and not at all misleading, and remains true going forward through time), this would be an extremely big deal. Modifications to the fundamental systems we use for collective communication and sense making is the obvious place a benevolent dictator would start to improve the current state of affairs.
However, history strongly suggests that this is not only not true, but most likely knowingly untrue (aka: a lie). I am obviously speculating, but I think this is a reasonable speculation.
Speculating sucks though. I think we are forced to do it far more often than is necessary (under the limitations imposed upon us by nature).
So how about this idea:
Let's say, in the spirit of uniting the country (USA), we make a bi-partisan decision to create a new role for the government: an as-honest-as-possible, process of constant audit of all "major" public communication platforms. Carefully selected, proven to be honest and trustworthy bi-partisan technical people (from the "grassroots" community) would be forcibly embedded within all major corporations, with 100% visibility into all source code, processes, and meetings (where "necessary"). They would carefully monitor the nature of all of this software that is exerting such a powerful force on our society, and that of the world. Where possible (which should be most of the time), their findings would be published for the public to see (and "sniff for imperfections or corruption").
These would be positions of extreme power and insight, and would offer genuine risks to intellectual property and confidential strategy of the companies subject to this treatment. As a first pass at managing this, these people could be paid extremely well, but they would also be subject to extremely punitive measures if they were to ever behave in a compromising manner.
Of course, the flaws in such a plan are numerous. Reality is complex - we can face that head on and manage it, or bury our heads in the sand with speculative claims like "...
> As we develop algorithms that power Birdwatch — such as reputation and consensus systems — we aim to(!) publish that code(!) publicly in the Birdwatch Guide. The initial ranking system for Birdwatch is already available here.
I think the idea of finding truth through consensus is the flaw here. A system like this would have silenced the great revolutionary thinkers of history.
> I think the idea of finding truth through consensus is the flaw here.
Oh, I'm not proposing that truth should be (or can be) defined by consensus, nor does the article.
This might be a decent example of the value of such a system: if someone makes such a proposal (ie: "Truth is defined by consensus."), or if someone asserts that someone else had made such a proposal, rather than the tweet being left as it is allowing the misinformation to propagate through the memeplex and into people's minds (aka: their reality), a formal rebuttal can be attached directly to it (so anyone happening upon can see that a collaborative consensus has been reached that it is plausibly misinformation), which is distinctly different from the tweet being declared(!) as conclusively(!) and unambiguously FALSE.
Sometimes such distinctions are important, sometimes they are not. In this case, the distinction is important, because of an anticipatory line I included in my comment:
> Of course, the flaws in such a plan are numerous. Reality is complex - we can face that head on and manage it, or bury our heads in the sand with speculative claims like "this wouldn't work".
In this case, you seem to be asserting that this wouldn't work, due to a specific flaw. However, the flaw you point out is not actually valid - I think this is a legitimate demonstration of how crowdsourcing can address a situation where the truth of a matter is "in play"...one person suggests an idea, another person asserts that it cannot work because <X>, another user realizes that <X> is a false statement and points that out. To be clear: this has not rendered the initial tweet to be True (that would be invalid logic), but it has cast significant and valid doubt on the assertion of ~"Viability = False" (because of a specific reason).
I think it wouldn't hurt to maybe have a little primer on the basics of logic embedded somewhere in this process as well - we always talk about how the general public needs improved critical thinking, so why not teach that wherever we can?
I think the general idea behind Birdwatch is plausibly workable and valuable, and using crowdsourcing allows it to scale. Similarly, I believe the general idea behind my idea to ensure that everyone is kept honest is also plausibly workable and valuable (until someone points out a valid reason that it isn't, of course).
> A system like this would have silenced the great revolutionary thinkers of history.
Can you explain your reasoning? I don't see anything in the proposal that suggests an intention to use this for precision or wholesale censorship - in fact, the article explicitly says the exact opposite of your concern:
> Eventually we aim to make notes visible directly on Tweets for the global Twitter audience, when there is consensus from a broad and diverse set of contributors.
> In this first phase of the pilot, notes will only be visible on a separate Birdwatch site. On this site, pilot participants can also rate the helpfulness of notes added by other contributors. These notes are being intentionally kept separate from Twitter for now, while we build Birdwatch and gain confidence that it produces context people find helpful and appropriate. Additionally, notes will not have an effect on the way people see Tweets or our system recommendations.
Are you seeing something here regarding censorship that I'm not?
That’s a pithy retort devoid of content. If 60% of the population believes the truth (for the sake of argument) and 40% believes another, the 40% can still do quite a bit of damage by intentionally spreading misinformation (causing confusion and chaos that the 60% have to distract themselves from) and/or violence (government overthrow doesn’t have to come from the majority).
Similarly, it’s not hard to imagine that if the minority is the one that’s correct, the majority can cause some damage enforcing the belief that the emperor is indeed wearing clothes.
The truth is the thing that needs protection, not speech. The challenge is that objective truth can be epistemologically difficult to protect without protecting lies (or conversely fighting falsehoods can inadvertently end up fighting some truths).
Worse, if the 40% are more enthusiastic about spreading their viewpoint than the 60% are about spreading the truth, then social media may have a majority of posts supporting the falsehood.
>The challenge is that objective truth can be epistemologically difficult to protect without protecting lies
Which is why even lies are, and should be, protected speech, even if that's not the current trendy thought.
Ultimately what it comes down to is: protecting unpopular minority opinions. If they're not protected, how would it even be possible to evaluate the "truth" of the content.
There is no universal truth. That's why everyone's speech should be protected. Even lies.
Consider: Every civil rights achievement made in the last 100 years. If the minority opinion was squelched without a chance to be considered, what progress would have been made?
Would a "Ministry of Truth" in whatever form it might take, have even allowed statements questioning racial or gender equality?
Any system that enforces such censorship becomes a theocracy dictating dogma.
A statement as simple as "There is no god" would become blasphemous again.
Actually I want 0 protections for lies. Whether the truth is in the majority or minority wouldn’t matter. The challenge is how to distinguish lies from truth. For what it’s worth the court system tends to be a place where that actually holds (discovered lies have a penalty). I think something similar could be done in the public square for public personalities as those personalities have an outsized influence on the public. In other words, as a public figure if you make blatantly false statements you open yourself to law suits where you have to demonstrate you had a good faith reason for your statement. This would instantly quash a lot of bullshit that politicians and celebrities spew with minimal downsides.
It’s the same line of reasoning that prevents public figures from suing people who make fun of them - when you’re a public figure you have an outsized responsibility to the public good.
To be fair, this is the same way all broadcast media has been since the Gutenberg Press was primarily used to print indulgences for the church. We got a good 25 years between the time the internet was relatively widespread before it was completely subverted by the powers that be, that’s pretty good. Now onward to the next communication technology, because the internet is now only useful for maintaining the status quo, just like TV, radio, newspaper, and every broadcast medium prior.
It don't think it is possible with a format like Twitter - I can imagine something like Wikipedia, but with sources limited to scientific publications, where each person improves upon a shared document.
Maybe Google wave is a better fit, allowing rich non-text items to enter it.
On Twitter you will have each person fight for their side, and you will never see a convergent product.
Of course this answer (and your question) presupposes that we are trying to achieve some form of shared consensus, whereas on Twitter, it is really two or more sides of a culture war fighting and wars are zero or negative sum games. If we are trying to find the truth I benefit even if every theory I had was wrong, because it still helped us get to the one that works, but in a war if I don't win I am sure to lose.
Introducing Watchbird, a community-based approach to misinformation (heh).
Posters: Get paid to post online, starting at $0.20/post! Our top posters earn up to $50/hour! Join now!
Sponsors: We have over 50,000 active users ready to post whatever you need online, no questions asked!
Similar services actually exist, but you know, one could always create a new one aimed at "fact checking" the fact checkers.
I guess my point is, you can't solve this problem with even more crowd bullshit. It needs to be done at a fundamental level, preferably by governments in school.
Afaik, there's still zero official classes/courses/lessons/whatever in most schools that would teach you to not trust everything you read and triple check everything yourself before believing anything.
Plus, this is pretty prone to abuse. Individuals are inherently dangerous, and crowds even more so. Someone doesn't like person X, so they "fact check" his tweets. Others see this coming from a "reputable" poster and jump on the bandwagon.
Seen it so many times it got old. Experimented with it myself. A post on Reddit (same content) that gets 8-12 fake upvotes in the first 30 minutes after being posted is infinitely more likely to start getting upvoted by hundreds of real users and get to the subreddit's hot front page than a post that got 0-3 upvotes, for example.
I was interested in Reddit's voting system and learned some interesting stuff. They're really smart about it, you can't just have multiple accounts and some proxies/VPNs and go at it like in the good old days. Votes are going to be ignored unless you know what you're doing. Probably not news for anyone in the industry, but I found it interesting.
Nearly all attempts to "fix misinformation" on social media I've seen in the last several years ranged from hopelessly clueless to downright sinister.
"Birdwatch allows people to identify information in Tweets they believe is misleading and write notes that provide informative context."
You're not going to fix anything by attacking the symptoms, which is exactly what this seems to propose. To fix the actual problem we need to create systems that generate and propagate trustworthy information, which people actually want to consume rather than attacking information you don't want people to consume.
"we have conducted more than 100 qualitative interviews with individuals across the political spectrum who use Twitter"
There is already a selection bias in play then, because there are large numbers of people who don't use Twitter for various reasons.
I agree with your characterization about attempts to fix misinformation. However, I don't think "what people actually want to consume" is a good signal for "trustworthy content"-if the social media era has shown us anything, it's that (barring a narrow band of critical, independent thinkers) we have an insatiable appetite for any content that reaffirms our tribalist priors, no matter how patently absurd the content may be.
To the extent that Twitter's goal is to combat post-truth-ism (as opposed to propagating it by way of its current business model), I think it pretty much has to develop its own strict code of ethics which prizes honesty, integrity, neutrality, and objectivity like the academics and journalists of yore. Specifically, we need a media landscape (including social media) that once again rewards both sides for bringing their best arguments, rather than dual partisan media outlets.
There's a very popular straw man counter-argument which is that I'm assuming that abhorrent racist ideas should be given the same attention as, I don't know, climate science, but that's not the case at all. Rather, I'm arguing that a neutralist and objective framework would discourage abhorrent racist ideas and encourage more respectable conservative intellectualism instead of the status quo which is to regard anything to the right of the far-left as uniformly vile (I'm oversimplifying a bit for sake of brevity). Similarly, such a platform wouldn't regard climate science and blank slatism as uniformly virtuous. The best left-wing thought would face-off against the best right-wing thought, and as a moderate liberal, I think the best left-wing thought will win.
Importantly, this pulls everyone toward more fact-based positions, which tends to have a moderating effect. Unfortunately, the corollary is that this presents a political obstacle--those who are deeply committed to these kinds of post-truth ideologies tend to vigorously oppose such reforms.
> [...] we’re designing Birdwatch to encourage contributions from people with diverse perspectives, and to reward contributions that are found helpful by a wide range of people.
> For example, rather than ranking and selecting top notes by a simple majority vote, Birdwatch can consider how diverse a note’s set of ratings is and determine whether additional inputs are needed before a consensus is reached. Additionally, Birdwatch can proactively seek ratings from contributors who are likely to provide a different perspective based on their previous ratings.
> Further, we plan for Birdwatch to have a reputation system in which one earns reputation for contributions that people from a wide range of perspectives find helpful.
I just want to point out that Minds[1] has a pretty clever content moderation policy that involves an appeal process with randomly selected users of the platform making blind judgements.
I haven’t been part of the process myself, nor have I used the platform yet at all. But this feature sounds quite good in theory.
A lot of people are wondering how this will stop misinformation. I agree that we can't crowdsource truth. But we can crowdsource information that can help reduce misinformation. When you have two sides disagreeing the first step is to build some common ground.
Twitter is trying to solve a tough problem. On one hand you've got people accusing Twitter of hosting and platforming hateful, harmful content. On the other hand you have people claiming that Twitter is calling the shots about what's true and suppressing information it doesn't like.
Maybe this is the first step towards something like a digital court. People on both sides present evidence, experts, witnesses. The two sides get a hand in picking the jury.
Or maybe the solvable problem is that information gets misconstrued and propagated. A video clip might get edited a certain way, for example. Solving this problem may not help us all agree on what happened in the video clip. However, we should at least be able to agree on what the two interpretations are. To make this happen, both sides would have to steel man the other side. Otherwise, the opposing side would claim they're being misportrayed. Having things that opposing sides agree upon would greatly help reduce unnecessary conflict.
There's a lot of reason to doubt that this will work. But one thing makes me hopeful that this might actually work is that Wikipedia's "Talk" pages appear to serve a similar purpose, and they serve that purpose adequately.
Sounds interesting, and I coyld definitely use something better than my filed, chaotic bookmarks to organize my collection of websites. The problem with this though? The same one that so many other systems for long term data organization can have: How do I know your service won't just die or fold sooner or later, leaving me and my organization efforts swinging in the wind? Bookmarks at least are just HTML files that can be exported, saved and imported across all browsers regardless of platform. Simple, resilient and thus robust.
81 comments
[ 5.6 ms ] story [ 134 ms ] threadAlthough Twitter's problem is way harder, IMO.
[0] https://www.youtube.com/watch?v=ObhK8lUfIlc
-Overwatch at least had some method of selecting people that "knew about the game." People that had never ranked competitively in CSGO couldn't make decisions if someone was cheating, whereas it seems that the hopes for this new Twitter feature will be that anyone can "fact-check" a tweet. Even if evidence is required, it is beyond easy to find secondary sources that skew facts or statistics into a different connotation.
-Whether or not someone is cheating is pretty binary. They either are or they aren't. One of my biggest concerns for Birdwatch is that it was likely be used on tweets that aren't binary statements of "write/wrong" facts, and will likely be used in mass-reports of those of other political stances. That, or shitposting
We need to see when one side ignores the other side. We need a list of unanswered questions to hold every side accountable.
Like you're going to hold the Leninists accountable for their unanswered claims against the Breadtubers for today's Voash drama?
Or intervene with a fact check when someone says the 8052 is UV reprogrammable when only the modern clones are?
Or for when the guy who talks about Magic the Gathering and was in Mulan makes an offhand comment about the new Harry Potter series and literally everyone gets upset?
How are you even dividing sides?
On the one hand, I certainly struggle to see how more crowds will improve upon what is fundamentally a problem with crowds to begin with.
On the other, Wikipedia seems to work reasonably well--and purely as a result of the community mores, not due to any sophisticated moderation algorithm or ranking structure or similar.
I'm not holding my breath, but it is interesting to see if a community-driven effort imbued with a Wikipedian-like spirit might succeed where more automated efforts have not.
Once your community or protocol "philosophically" rejects checksums it is only a matter of time before your system breaks catastrophically.
This philosophical freedom vs authoritarianism thing you're proposing is a view that is incomplete and small to the point of being false.
The actual underlying question is do you accept that grounded truth exists and are you willing to have a large and nuanced enough view to include all non false propositions and reject all false propositions.
But if you're squinting that hard, not a lot of light is getting through.
When you only have a hammer, every problem looks like a nail.
I believe that the social networks have built technologies that probably can't actually be made to work in a socially beneficial way, since they're utterly dependent on crowdsourcing and automation for economic reasons, but those tools aren't actually up the the problems they face.
The answer to "who fact-checks the fact-checkers" is "the fact checkers".
And we can use consensus, reputation, and checks against known facts to validate the checks themselves. If group A, group B, and group C all agree and group D disagrees, then we can say with relative confidence that D is probably wrong in this case. It is safe to go with the consensus.
However, if A, B, C are notoriously unreliable and have a history of distorting claims, then it is beneficial to see why D disagrees with all three. Or even not use A, B, and C as sources.
We can also judge a fact-checker's veracity by how they evaluate a claim. Do they provide sources for their claim. It's one thing to say "This will murder 40 people a minute" and quite another to say "This claim disagrees with a recent study by the CDC [link to study] that says it will only murder 39 people a minute".
The only reason to cast doubt on the process or activity of fact-checking is if you don't want people checking your facts.
How much does truth matter in a post-truth society?
People have spent enormous sums of money and time to hide the truth from the public. Just because every effort has been made to hide the truth, doesn’t diminish the fact that the truth matters. In my opinion, it’s just the opposite. It matters more than ever.
That kind of depends if you are discussing something objective or not. Wikipedia seems to have developed a strong left wing bias on anything political these days. Articles without any kind of politics still seem to be pretty objective. Can there be a single truth in politics, where ideas are very subjective?
The truth still reigns supreme, it is just that people sometimes suffer unnecessarily because they do not know what the truth is.
As a specific example, as a nonbinary person, I'm constantly running into people online who tell me there are only two genders, or that singular they is some sort of new concept. My concern with a consensus system is that it will be used to shut down people like me (or trans folk, or socialists, etc. etc. etc.)
How do you build a consensus system that protects minority persons and also weeds out misinformation?
Edit: Downvotes for... being non-binary I guess?
Gender, in the context of language referring to a person, has a fairly settled meaning.
https://www.oed.com/viewdictionaryentry/Entry/77468
This is doubly true for safety systems. I don't see much of the press release focused on how abuse of the safety system will be avoided beyond, "We'll eventually let it be reviewed by outside parties."
What's to stop a flood of Russian accounts gaining control of what's marked as disinformation?
Consensus is the enemy of understanding. For topics where there is conflicting or poor evidence or that bear on the culture war, I do not want people voting on what is the consensus truth. I want to see all the evidence. We have enough problems with researchers not publishing uncomfortable data; I don't want the little that exists flagged because it conflicts with the average Twitter user's sensibilities.
That's a nice sentiment in a vacuum, but we're not in a vacuum. It's sort of like focusing on getting every child a college education when you can't even manage to teach them all literacy yet.
At the point, I feel like the priority needs to be around figuring out a way to defeat disinformation and misinformation to restore some kind of common ground understanding of the facts. The marketplace of ideas can't function if major factions reject truth-seeking in favor of blatant lies that are appealing for various other reasons (e.g. emotional satisfaction, ability to manipulate others, etc.).
You're confusing preferences and facts. I was talking about facts [1], you're talking about preferences.
[1] like Donald Trump is not waging a secret war to defeat a conspiracy of Satan-worshiping pedophiles who harvest a (fictitious) drug from trafficked children and Trump lost the 2020 election fairly.
The problem always becomes that few think their faction has this problem. Most people think that they and their side believe in the truth, and that the other side is peddling false information. In such an environment, "combating misinformation" inevitably becomes "trying to get everyone to agree with me."
In order to combat misinformation we'd probably have to start by combating the emotional commitment we hold to a certain narrative before we even examine all of the facts. That's not easy to do, but I imagine a good first step would be to cut out the constant 24/7 news cycle that a lot of people are addicted to and that seems to feed into these emotions. Twitter, of course, is a large source of this stuff.
Getting away from the news is an excellent way to start thinking more clearly.
The problem with your line of thinking is that it assumes good faith and a certain amount of competence, which are assumptions that I don't think we can reasonably make anymore in light of the insane success of things like QAnon and "Stop the Steal."
Conservative misinformation is a big talking point for liberals, and maybe the big societal issue at the moment, but as a small scale test run, I'd love to see birdwatch try to correct the record for misinformation that is commonly believed in liberal circles: Anti-GMO, anti-vaxx, toxic whiteness, the extents of systemic racism in our society, some of the more dire prognostications of nuclear war and global warming.
As it's becoming an ever-increasing method of communication between government officials and whatnot, I'd say some amount of "truthfulness" and "validity" is warranted. The other day a member of the US government (I think it was Ted Cruz? I can't find it at the moment) wrote a tweet that implied he thought the Paris Climate agreement only affected the citizens of Paris. Something as factually "write/wrong" as that should be able to be contested and held accountable to being "wrong"
That being said, I agree with you that this will not be used as intended and will undoubtedly lead to more confusion and confirmation bias from all sides. Using crowd-sourced knowledge in this manner leads to the assumption that "what the majority thinks is right is factual," and I'm reminded it was once "fact" that the sun rotated around the Earth
Any tweets questioning the official narrative will be roundly criticized and ridiculed. Instead of having to delve into a long reply thread to see debunkers make their case, there will be an easy to digest notice within grasp. Which is not necessarily a bad thing if it is used fairly and responsibly. But judging on Twitter’s past performance, it likely won’t be.
Dissent will be publicly humiliated while pure disinformation from governments, think tanks, and corporations have no such objections. Any doubts to the accepted story will be pointed at one of the fact checker sites and further inquiries censored.
In an honest world, this would be one giant step to finally getting at the truth. But this isn’t an honest world, is it?
Not to mention that in an honest wold, why would you even need something like this?
You don't need it in an honest world, and it will end up being worse than nothing in a dishonest world. It doesn't seem like a good idea in any scenario.
The world is what we make it - we are the masters of our own destiny. This is literally true (subject to boundaries imposed on us by nature: the laws of physics, the physical resources available to us, the nature of the evolved human mind and the societies we have built for ourselves to operate in, and other things I may overlook).
If a variable in the world is not to our liking, and it is fixable, we can choose to fix it (as a collective species), or not. Mother Nature imposes some restrictions on this, but not many.
From the article:
> To date, we have conducted more than 100 qualitative interviews with individuals across the political spectrum who use Twitter, and we received broad general support for Birdwatch. In particular, people valued notes being in the community’s voice (rather than that of Twitter or a central authority) and appreciated that notes provided useful context to help them better understand and evaluate a Tweet (rather than focusing on labeling content as “true” or “false”). Our goal(!) is to build Birdwatch in the open, and have it shaped(!) by the Twitter community.
> To that end, we’re also taking significant steps to make Birdwatch transparent:
- All(!???) data contributed to Birdwatch will be publicly available and downloadable in TSV files
- As we develop algorithms that power Birdwatch — such as reputation and consensus systems — we aim to(!) publish that code(!) publicly in the Birdwatch Guide. The initial ranking system for Birdwatch is already available here.
> We hope this will enable experts, researchers, and the public to analyze or audit Birdwatch, identifying opportunities or flaws that can help us(!) more quickly build an effective(!) community-driven(!) solution.
If what they are saying is 100% true (and not at all misleading, and remains true going forward through time), this would be an extremely big deal. Modifications to the fundamental systems we use for collective communication and sense making is the obvious place a benevolent dictator would start to improve the current state of affairs.
However, history strongly suggests that this is not only not true, but most likely knowingly untrue (aka: a lie). I am obviously speculating, but I think this is a reasonable speculation.
Speculating sucks though. I think we are forced to do it far more often than is necessary (under the limitations imposed upon us by nature).
So how about this idea:
Let's say, in the spirit of uniting the country (USA), we make a bi-partisan decision to create a new role for the government: an as-honest-as-possible, process of constant audit of all "major" public communication platforms. Carefully selected, proven to be honest and trustworthy bi-partisan technical people (from the "grassroots" community) would be forcibly embedded within all major corporations, with 100% visibility into all source code, processes, and meetings (where "necessary"). They would carefully monitor the nature of all of this software that is exerting such a powerful force on our society, and that of the world. Where possible (which should be most of the time), their findings would be published for the public to see (and "sniff for imperfections or corruption").
These would be positions of extreme power and insight, and would offer genuine risks to intellectual property and confidential strategy of the companies subject to this treatment. As a first pass at managing this, these people could be paid extremely well, but they would also be subject to extremely punitive measures if they were to ever behave in a compromising manner.
Of course, the flaws in such a plan are numerous. Reality is complex - we can face that head on and manage it, or bury our heads in the sand with speculative claims like "...
I think the idea of finding truth through consensus is the flaw here. A system like this would have silenced the great revolutionary thinkers of history.
Oh, I'm not proposing that truth should be (or can be) defined by consensus, nor does the article.
This might be a decent example of the value of such a system: if someone makes such a proposal (ie: "Truth is defined by consensus."), or if someone asserts that someone else had made such a proposal, rather than the tweet being left as it is allowing the misinformation to propagate through the memeplex and into people's minds (aka: their reality), a formal rebuttal can be attached directly to it (so anyone happening upon can see that a collaborative consensus has been reached that it is plausibly misinformation), which is distinctly different from the tweet being declared(!) as conclusively(!) and unambiguously FALSE.
Sometimes such distinctions are important, sometimes they are not. In this case, the distinction is important, because of an anticipatory line I included in my comment:
> Of course, the flaws in such a plan are numerous. Reality is complex - we can face that head on and manage it, or bury our heads in the sand with speculative claims like "this wouldn't work".
In this case, you seem to be asserting that this wouldn't work, due to a specific flaw. However, the flaw you point out is not actually valid - I think this is a legitimate demonstration of how crowdsourcing can address a situation where the truth of a matter is "in play"...one person suggests an idea, another person asserts that it cannot work because <X>, another user realizes that <X> is a false statement and points that out. To be clear: this has not rendered the initial tweet to be True (that would be invalid logic), but it has cast significant and valid doubt on the assertion of ~"Viability = False" (because of a specific reason).
I think it wouldn't hurt to maybe have a little primer on the basics of logic embedded somewhere in this process as well - we always talk about how the general public needs improved critical thinking, so why not teach that wherever we can?
I think the general idea behind Birdwatch is plausibly workable and valuable, and using crowdsourcing allows it to scale. Similarly, I believe the general idea behind my idea to ensure that everyone is kept honest is also plausibly workable and valuable (until someone points out a valid reason that it isn't, of course).
> A system like this would have silenced the great revolutionary thinkers of history.
Can you explain your reasoning? I don't see anything in the proposal that suggests an intention to use this for precision or wholesale censorship - in fact, the article explicitly says the exact opposite of your concern:
> Eventually we aim to make notes visible directly on Tweets for the global Twitter audience, when there is consensus from a broad and diverse set of contributors.
> In this first phase of the pilot, notes will only be visible on a separate Birdwatch site. On this site, pilot participants can also rate the helpfulness of notes added by other contributors. These notes are being intentionally kept separate from Twitter for now, while we build Birdwatch and gain confidence that it produces context people find helpful and appropriate. Additionally, notes will not have an effect on the way people see Tweets or our system recommendations.
Are you seeing something here regarding censorship that I'm not?
Similarly, it’s not hard to imagine that if the minority is the one that’s correct, the majority can cause some damage enforcing the belief that the emperor is indeed wearing clothes.
The truth is the thing that needs protection, not speech. The challenge is that objective truth can be epistemologically difficult to protect without protecting lies (or conversely fighting falsehoods can inadvertently end up fighting some truths).
Which is why even lies are, and should be, protected speech, even if that's not the current trendy thought.
Ultimately what it comes down to is: protecting unpopular minority opinions. If they're not protected, how would it even be possible to evaluate the "truth" of the content.
There is no universal truth. That's why everyone's speech should be protected. Even lies.
Consider: Every civil rights achievement made in the last 100 years. If the minority opinion was squelched without a chance to be considered, what progress would have been made?
Would a "Ministry of Truth" in whatever form it might take, have even allowed statements questioning racial or gender equality?
Any system that enforces such censorship becomes a theocracy dictating dogma.
A statement as simple as "There is no god" would become blasphemous again.
It’s the same line of reasoning that prevents public figures from suing people who make fun of them - when you’re a public figure you have an outsized responsibility to the public good.
Maybe Google wave is a better fit, allowing rich non-text items to enter it.
On Twitter you will have each person fight for their side, and you will never see a convergent product.
Of course this answer (and your question) presupposes that we are trying to achieve some form of shared consensus, whereas on Twitter, it is really two or more sides of a culture war fighting and wars are zero or negative sum games. If we are trying to find the truth I benefit even if every theory I had was wrong, because it still helped us get to the one that works, but in a war if I don't win I am sure to lose.
This seems to be the heart of the problem.
Posters: Get paid to post online, starting at $0.20/post! Our top posters earn up to $50/hour! Join now!
Sponsors: We have over 50,000 active users ready to post whatever you need online, no questions asked!
Similar services actually exist, but you know, one could always create a new one aimed at "fact checking" the fact checkers.
I guess my point is, you can't solve this problem with even more crowd bullshit. It needs to be done at a fundamental level, preferably by governments in school.
Afaik, there's still zero official classes/courses/lessons/whatever in most schools that would teach you to not trust everything you read and triple check everything yourself before believing anything.
Plus, this is pretty prone to abuse. Individuals are inherently dangerous, and crowds even more so. Someone doesn't like person X, so they "fact check" his tweets. Others see this coming from a "reputable" poster and jump on the bandwagon.
Seen it so many times it got old. Experimented with it myself. A post on Reddit (same content) that gets 8-12 fake upvotes in the first 30 minutes after being posted is infinitely more likely to start getting upvoted by hundreds of real users and get to the subreddit's hot front page than a post that got 0-3 upvotes, for example.
I was interested in Reddit's voting system and learned some interesting stuff. They're really smart about it, you can't just have multiple accounts and some proxies/VPNs and go at it like in the good old days. Votes are going to be ignored unless you know what you're doing. Probably not news for anyone in the industry, but I found it interesting.
Whatever the Twitter fact checking note says, you should still not take any information from the internet at facee value.
"Birdwatch allows people to identify information in Tweets they believe is misleading and write notes that provide informative context."
You're not going to fix anything by attacking the symptoms, which is exactly what this seems to propose. To fix the actual problem we need to create systems that generate and propagate trustworthy information, which people actually want to consume rather than attacking information you don't want people to consume.
"we have conducted more than 100 qualitative interviews with individuals across the political spectrum who use Twitter"
There is already a selection bias in play then, because there are large numbers of people who don't use Twitter for various reasons.
To the extent that Twitter's goal is to combat post-truth-ism (as opposed to propagating it by way of its current business model), I think it pretty much has to develop its own strict code of ethics which prizes honesty, integrity, neutrality, and objectivity like the academics and journalists of yore. Specifically, we need a media landscape (including social media) that once again rewards both sides for bringing their best arguments, rather than dual partisan media outlets.
There's a very popular straw man counter-argument which is that I'm assuming that abhorrent racist ideas should be given the same attention as, I don't know, climate science, but that's not the case at all. Rather, I'm arguing that a neutralist and objective framework would discourage abhorrent racist ideas and encourage more respectable conservative intellectualism instead of the status quo which is to regard anything to the right of the far-left as uniformly vile (I'm oversimplifying a bit for sake of brevity). Similarly, such a platform wouldn't regard climate science and blank slatism as uniformly virtuous. The best left-wing thought would face-off against the best right-wing thought, and as a moderate liberal, I think the best left-wing thought will win.
Importantly, this pulls everyone toward more fact-based positions, which tends to have a moderating effect. Unfortunately, the corollary is that this presents a political obstacle--those who are deeply committed to these kinds of post-truth ideologies tend to vigorously oppose such reforms.
> For example, rather than ranking and selecting top notes by a simple majority vote, Birdwatch can consider how diverse a note’s set of ratings is and determine whether additional inputs are needed before a consensus is reached. Additionally, Birdwatch can proactively seek ratings from contributors who are likely to provide a different perspective based on their previous ratings.
> Further, we plan for Birdwatch to have a reputation system in which one earns reputation for contributions that people from a wide range of perspectives find helpful.
https://twitter.github.io/birdwatch/about/challenges/
I haven’t been part of the process myself, nor have I used the platform yet at all. But this feature sounds quite good in theory.
[1]: https://www.minds.com/content-policy
Twitter is trying to solve a tough problem. On one hand you've got people accusing Twitter of hosting and platforming hateful, harmful content. On the other hand you have people claiming that Twitter is calling the shots about what's true and suppressing information it doesn't like.
Maybe this is the first step towards something like a digital court. People on both sides present evidence, experts, witnesses. The two sides get a hand in picking the jury.
Or maybe the solvable problem is that information gets misconstrued and propagated. A video clip might get edited a certain way, for example. Solving this problem may not help us all agree on what happened in the video clip. However, we should at least be able to agree on what the two interpretations are. To make this happen, both sides would have to steel man the other side. Otherwise, the opposing side would claim they're being misportrayed. Having things that opposing sides agree upon would greatly help reduce unnecessary conflict.
Over a twitter spat? Seems like a waste of resources.