Long rant, but I feel it's necessary. Part of the value that fact-checkers provide is that they often give context to a piece of misinformation or lie.
For example, this statement, which I assume is coming from ChatGPT:
"3. The Steele Dossier and Trump-Russia Collusion (2016–2019)
Initial Claim: Many outlets and fact-checkers treated claims from the Steele Dossier about Trump’s connections to Russia as credible and verified.
Reassessment: Subsequent investigations, including the Mueller Report and the Durham Report, found no evidence of collusion and discredited many of the dossier’s claims.
Criticism: Media outlets were criticized for amplifying unverified information, leading to years of political and public mistrust."
This seems just flat wrong. Why? Because the Muller team "didn't evaluate 'collusion' with the Russian government" during the course of its investigation. It simply wasn't a thing they looked into because collusion isn't a legal term (https://www.politico.eu/article/mueller-refutes-trumps-no-collusion-no-obstruction-line/). That context — which is crucial to understanding the story — is missing.
But even the premise of the claim — that "Many outlets and fact-checkers treated claims from the Steele Dossier about Trump’s connections to Russia as credible and verified," — is misleading. Again, the Mueller team didn't include the Steele Dossier in their investigation because they conducted their own investigation with primary sources.
The Mueller team didn't discredit the dossier because it wasn't in the scope of their task. This is akin to asking why a line chef didn't personally deliver a dish to a table. As for media outlets, that's how a news story evolves throughout its lifecycle (more on that in a minute).
The last line, "Media outlets were criticized for amplifying unverified information, leading to years of political and public mistrust," is just childish. Of course news organizations often run with unverified information in their stories. They even get facts wrong in their initial coverage!
This is because not all information is available to a journalist at the outset of a story. You have an allegation. You investigate. You publish that story and you keep investigating and publishing additional stories as the investigation evolves. You correct the story as new information is introduced. That's how news works. It's hard! And it's made even more difficult when there are a lot of competing interests in suppressing or spinning information — such as almost literally anything related to the Trump administration.
At this moment, I have two central issues with removing third-party fact-checkers. First, it places the onus of fact-checking on a community that's inherently media illiterate (https://www.cnn.com/2024/01/21/us/teaching-media-literacy-in-schools/index.html), which can easily lead to a sustained spike in disinformation.
This leads me to a tangential issue: social media networks are built to be manipulated since they're built for engagement over veracity. Facebook and Twitter make money off of selling ads — not by facilitating honest conversation. How can you easily sell more ads? By having your algorithms favor posts and media that drives engagement over the mundane truth. Which is sad.
I guess, in sum, this is a bad but predictable move by Facebook. It does nothing beneficial for its users while offering a fig leaf towards gaming engagement via algorithms.
ZEBLEY: Well, in that instance, we said we did not find sufficient evidence of a conspiracy to charge a conspiracy. Collusion is not really a criminal violation. So we applied that term to mean conspiracy. And we ultimately said, look, we didn't find sufficient evidence to say there was a conspiracy. But we did say that the Russians interfered in the election with the intention of favoring Donald Trump because they perceived there'd be a benefit from a Trump presidency, and they interfered in a way to harm the other candidate, Hillary Clinton. And we also found evidence that the Trump campaign perceived it would benefit from some of the Russian conduct. For instance, they perceived they would benefit from the releases by Wikileaks, and at the time, Wikileaks was releasing information that the Russians had stolen from the Clinton campaign.
GROSS: There was also the infamous Trump Tower meeting where a Russian proposed to Don Jr. that this Russian had damaging information about Hillary Clinton that the Trump campaign could use and offered to have a meeting to hear about it. And Don Jr. agreed to do it. And so Don Jr., Eric Trump, and Paul Manafort, who was the campaign manager at the time, met with a couple of Russians. As it turns out, the Russians had their own agenda, but the fact that the Trump sons and campaign manager were willing to have that meeting, in terms of information, what did that tell you?
These were just a few examples from a much longer piece on this topic summarized - I figured someone would comment on one here, which is great bc it proves the point - that a community (of smart people like you, whose ideas I respect & appreciate) can add input to things, which just happened. I would just rather have this effort done in a decentralized fashion by many citizens than a few people (who frequently have specific political agendas).
I hesitate to take Meta at face value on anything, but I agree that the fact-checking industry has been distorted and overextended from the very specific (and valuable) role it's long had in journalism and other publishing. And it's quite possible the downstream effects of this reputation harm are worse than if there had never been high-profile fact-checking operations.
I'd also argue that Biden's cognitive health (or Trump's in his 1st term) is an area where fact-checking isn't well-suited. It's one thing to accurately convey official doctor's records, etc., as is always done with presidents, but the armchair diagnoses over the past 8 years have been guesses at best, not factual analyses.
As for the inflation and COVID origins situations, those illustrate how fact-checking is used prematurely for ongoing situations. It's one thing to say, for instance, "So-and-so claimed China did COVID, of which there's not proof. Here is what we know about the possibilities." But too many of these went further, as you note, as if there wasn't more to learn.
Great post Adam. I'm a massive fan of Community Notes on X. IMO it is the single most important feature created by any social media company since the invention of the newsfeed. CN should be on the front page of ALL media outlets (social or publishing).
An unappreciated reason why it succeeds is that it's anonymous. A CN must stand on its own merits, independent of who wrote it. Likewise anonymity means that there's no social status to be gained within the community of Note writers. Compare this to Wikipedia where mods entire identities are created around the social structure of editors.
And a final thought - why is it that we use secret ballots at elections? To avoid preference falsification.
The closer our institutions and platforms are to a meritocracy the better. Community notes and crowdsourcing are not 100% accurate or precise, but more times than not they correct narratives that are misleading or flat out false. No one is immune to propaganda, but this mechanism is a good counter to it or at the very least provides good counter points to prevailing arguments.
Once again, you're unfortunately missing the point. We have numerous studies that show fact checking is not biased, but that conservatives share significantly more misinformation than others. That is why they are fact checked more often! Pursuing political balance in fact checking actually makes it more political, not less.
And the claim that fact checkers in Texas are some how less biased than California?? That's absurd.
Community notes are a good step forward, but they are NOT a replacement for fact checking. While fact checkers can get it wrong, and community notes can get it right, that is not evidence that fact checkers are always wrong or that community notes are right more often!
As long as social media companies have algorithms that choose to promote some content, they are responsible for spreading misinformation. Calling that responsibility "censorship" is not a productive framing.
Thanks for the comments Sean I always appreciate your perspective. If you dig enough into politicians you'll see both sides lie, more resources are simply allocated to one side here so there is a lot of selection bias. If you get a large enough sample of politicians this has gotta be evenly distributed. Unsure how you'd test but if you listen to Congressional hearings on CSPAN etc unedited by TV news you get a sense of how there's reasonable dems and cons, it actually makes me way more optimistic for our institutions than cable news etc.
The culture in California is *definitely* super biased, I've lived in 4 major cities in America >2 years and can see that. Moving the teams to Texas was symbolic sure, but honestly I bet you get more centrism here. Anyway separately Zuck stated they're gonna focus trust and safety on other things. Because of this they will have *more* resources to tackle actually bad and illegal actors, predators, spammers and stalkers. This is what they should be doing, we all agree that stuff is bad.
Glad you agree community notes are cool. The algo discussion is an entirely other team than the in-house and outsourced censorship folk ofc, so a very different discussion. I agree with you here too those algos particularly on FB should bias to content from friends vs politics or viral slop. IDK why they do that.
This isn't a matter of opinions, we have real scientific studies that show conservatives spread more misinformation. Here is one, I suggest you read a few of them: https://www.nature.com/articles/s41586-024-07942-8
As for Texas being less biased than California, it really is absurd. Both states are enormous and have a multitude of people with many different beliefs. The idea that geographic location determines your bias is insane and evidence of opinions over facts. Your personal network is not representative of an entire state!
This isn't just a topic where opinions can differ, that's part of the problem. There are real facts here and you can't disagree on facts.
Texas has a more balanced mix of urban liberal areas and rural conservative areas, while California is overwhelmingly Democratic. I am not considering my personal network I'm looking at this at the macro level and fairly dispassionately. You're right though this shouldn't determine N=1 person's bias. It's symbolic anyway, FB just needs to hire better. You go to work to make revenue for the company, not participate in activism.
There is bias in these studies btw, how do we define misinformation? There's left leaning people arguing against 2+2 = 4 in academia and other areas. No one is without fault. Also using Twitter here presents additional selection bias, and a lot of what is in that link is outright spam. Regardless none of it changes that I don't think tech companies should be employing ideologically-driven fact checkers.
I believe you that they are your opinions, but we have facts on this. We can't dismiss studies because you think there is *hand waving* bias. That's exactly the problem that fact checkers need to address.
Texas is not less biased than California. Based on voting records, Texas is at least as conservative as California is liberal but even that is unreliable.
I appreciate we can have a friendly discussion about this, but it's evidence of why fact checking is necessary. There is truth here and opinions, and normal folks don't have the time to separate the two.
Appreciate it too. And look on opinion pieces those are just that, they're labeled as such or on an opinion blog and shouldn't need community notes so much, they're presenting ideas and narrative and in good cases have factual proof points/links too. Specific breaking news stories or comments and promises from elected officials *should* be held to account. We agree on this! A new experiment is about to unfold so we'll see what happens. I remain optimistic.
Sean, the main issue that I’ve had is that, on important issues, fact checking and censorship is one sided.
For instance, on Covid 19 vaccinations, I don’t recall anyone being fact checked in 2021 for saying that vaccines were “effective” at preventing transmission, but plenty were banned for (correctly) saying the opposite. The impact was that many people then self censored.
On climate change, I don’t see any media being fact checked for (incorrectly) saying that hurricanes have grown more frequent and/or more intense in the Atlantic basin. On the other hand, plenty have been censored for correctly presenting the scientific evidence. See Roger Pielke Jr’s post from yesterday.
Would you argue fact checking has been balanced on these two topics? If not, are there meaningful policy discussions in which fact checking is balanced?
You seem to have some anecdotes about fact checking like Adam does, and anecdotes are not data. I'm not going to debate what you recall or experienced, which sounds different from me.
Our discussion shows the limitations of fact checking and highlights the importance of robust, uncensored discussions.
About hurricanes — I would guess the article you link bases its claim on data collected from about 1980. There’s a valid reason to do that (we didn’t have accurate satellite imagery before). However the weakness is that hurricane activity shows strong decadal patterns, and 1980 was a low in hurricane activity. I’m basing my claim on NOAA and the IPCC AR6 WG1, which take into account data starting from the early 1900’s.
We can spend days discussing this with intellectual honesty and still end up disagreeing. In other words, this is an area where fact checking is not useful, and actually prevents this kind of necessary discussion.
Fact checking is useful for trivial statements like “there are 53 states in the US” or “the iPhone was invented in 1453,” not much else.
No, I linked to a study from last year that focuses on hurricanes in the last decade. What this demonstrates is the need for fact checking because you are convinced of an incorrect conclusion and haven't revisited your priors. You cannot rely on community notes to remedy those things, as all they end up doing is reinforcing popular priors.
Generally speaking, the need for fact checking is that social media allows they misconceptions to spread and become priors for other people. Just like you dismissed the study I shared, people tend to avoid uncomfortable truths and community notes doesn't address that at all.
Sean, this is a worse study than I was giving it credit for. “Focuses on hurricanes in the last decade”: there is no way that one can draw any conclusion on the relationship between global warming and increased hurricane intensity over such a short amount of time (10 years). That would be the case for any climate phenomenon but even more so for hurricane activity which shows strong auto correlation.
There’s one way to avoid cherry picking studies of dubious quality and that is to rely on assessments such as the IPCC.
Honestly, if you aren't going to read the studies then I'm going to leave the conversation here. It doesn't just study the last decade, but it focuses on whether the past decade fits historical patterns. It's a good study, and I recommend doing a lot more reading on the topic before you form conclusions.
I wonder if this will actually apply to topics like criticism of Israel, which Meta was recently found to be suppressing. It’s hard for me to believe Zucc won’t carve out exceptions.
Community notes is great, and is one of the few improvement made to the platform. I don't trust Zuck, nor the motives behind this, but, as you state, as long as Meta is still able to (or willing to) remove genuinely harmful, dangerous content, then supplementing that with community moderation seems a sound decision.
It's disingenuous to say that community notes will work to monitor and counter disinformation and hate speech when it hasn't worked on X. By the time community notes can submit counterfactuals, it's too late. Misinformation has already gone viral. Besides, people simply take screenshots of the post before it's flagged and propagate it this way. It's easy to get around community notes. Community notes sound good, but in practice, it doesn't work.
Your analysis was remiss to note the political context of Zuckerberg's decision. I conclude that Zuckerberg's decision to remove fact-checking is politically expedient for him to do so. Consider the fact that last week, Meta’s global policy chief, Nick Clegg — a former British deputy prime minister who was chosen for his centrist bona fides — was replaced by Joel Kaplan, a longtime Republican operative who has acted for years as Mr. Zuckerberg’s liaison to the pro-Trump right. On Monday, Meta announced the appointment of three new board members, including Dana White, the chief executive of the Ultimate Fighting Championship and a close friend and political ally of Mr. Trump’s.
Zuckerberg promotes his decision as a return to free speech ideals and the need to avoid too much censorship. However, he fails to acknowledge what had happened on Meta’s sites when misinformation or foreign propaganda was permitted to propagate unchecked.
Unfettered hate speech has real-world consequences. Take, for example, what happened in Myanmar in 2017, where the Rohingya were killed, tortured, raped, and displaced in the thousands as part of the Myanmar security forces’ campaign of ethnic cleansing. In the months and years leading up to the atrocities, Facebook’s algorithms were intensifying a storm of hatred against the Rohingya, which contributed to real-world violence. How are Community Notes supposed to combat this?
For those who want a counter-argument to Adam's defense of Zuckerberg's decision, I suggest you read The Big Picture's substack entitled "Royally Zucked": https://shorturl.at/kuocl.
Just want to note that unfettered hate speech is not even allowed on Twitter/X, their team removes it if it violates their guidelines (https://help.x.com/en/rules-and-policies/hateful-conduct-policy). Very safe to assume Facebook will still do the same. It really appears to me they are simply allowing discussion around a bunch of topics some people don't like. They even noted they are doubling down on removing grifters, spammers, predators etc which these resources should be devoted to.
Meta changed their hate speech guidelines for the worse. See Wired's article https://www.wired.com/story/meta-immigration-gender-policies-change/. One example: Under the old policy, users were prohibited from referring to women as "household objects or property.” That part of the former policy is now crossed out, so we can assume that the objectification of women is now fair game.
Of course, I'm not going to defend that. But I wonder to what extent certain policies like this would change much, were people really chomping at the bit to say certain things like this? Anyone who would harass you before you likely already blocked or muted.
It's a $1.5T company and I don't think they make these decisions lately. The tech press foams at the mouth to write negative stories here about any policy decisions that happen and will bias to be negative before things even occur. Of course, I personally bias to free speech (not to harass people, just to not shrink the spectrum of our ideas). That's probably why we post more here and not there, and with this there is very little bad behavior. Appreciate the thoughts.
Pretty easy to be on the right side of this one, if you're on the side of the censorship-industrial complex at least consider if you are with the bad guys. This post isn't political. Also, if you can't understand why an election was lost, well, there's probably no helping you.
The fact that all the misinformation researchers and fact checkers lean one way shows why we should wish the community to do this and not individuals. So that becomes inherently democratic and truth-seeking which is what America is.
Is it possible that one side is a lot more likely to spread misinformation though? While I think you make some good points on the flaws of a fact checking approach to Social media, I think you're completely ignoring the objective truth that there is a deluge of misinformation coming from one side. Yes there's misinformation from the left too, but volume and intensity wise it's not comparable.
Long rant, but I feel it's necessary. Part of the value that fact-checkers provide is that they often give context to a piece of misinformation or lie.
For example, this statement, which I assume is coming from ChatGPT:
"3. The Steele Dossier and Trump-Russia Collusion (2016–2019)
Initial Claim: Many outlets and fact-checkers treated claims from the Steele Dossier about Trump’s connections to Russia as credible and verified.
Reassessment: Subsequent investigations, including the Mueller Report and the Durham Report, found no evidence of collusion and discredited many of the dossier’s claims.
Criticism: Media outlets were criticized for amplifying unverified information, leading to years of political and public mistrust."
This seems just flat wrong. Why? Because the Muller team "didn't evaluate 'collusion' with the Russian government" during the course of its investigation. It simply wasn't a thing they looked into because collusion isn't a legal term (https://www.politico.eu/article/mueller-refutes-trumps-no-collusion-no-obstruction-line/). That context — which is crucial to understanding the story — is missing.
But even the premise of the claim — that "Many outlets and fact-checkers treated claims from the Steele Dossier about Trump’s connections to Russia as credible and verified," — is misleading. Again, the Mueller team didn't include the Steele Dossier in their investigation because they conducted their own investigation with primary sources.
The Mueller team didn't discredit the dossier because it wasn't in the scope of their task. This is akin to asking why a line chef didn't personally deliver a dish to a table. As for media outlets, that's how a news story evolves throughout its lifecycle (more on that in a minute).
The last line, "Media outlets were criticized for amplifying unverified information, leading to years of political and public mistrust," is just childish. Of course news organizations often run with unverified information in their stories. They even get facts wrong in their initial coverage!
This is because not all information is available to a journalist at the outset of a story. You have an allegation. You investigate. You publish that story and you keep investigating and publishing additional stories as the investigation evolves. You correct the story as new information is introduced. That's how news works. It's hard! And it's made even more difficult when there are a lot of competing interests in suppressing or spinning information — such as almost literally anything related to the Trump administration.
At this moment, I have two central issues with removing third-party fact-checkers. First, it places the onus of fact-checking on a community that's inherently media illiterate (https://www.cnn.com/2024/01/21/us/teaching-media-literacy-in-schools/index.html), which can easily lead to a sustained spike in disinformation.
Two, the trust that community notes add up to anything is entirely reliant on the social media network in the first place (see: https://www.washingtonpost.com/technology/2024/10/30/elon-musk-x-fact-check-community-notes-misinformation/ AND https://www.forbes.com/sites/antoniopequenoiv/2023/07/25/twitter-deletes-fact-check-of-musk-connecting-bronny-james-cardiac-arrest-to-covid-vaccine/). If someone like Elon Musk, who has a vested interest in promoting a partisan narrative over verifiable information (read abstract: https://www.pnas.org/doi/10.1073/pnas.2025334119), then how is a casual user to get information from a reliable source in the first place? And, crucially, isn't weighing an algorithm a type of censorship to begin with?
This leads me to a tangential issue: social media networks are built to be manipulated since they're built for engagement over veracity. Facebook and Twitter make money off of selling ads — not by facilitating honest conversation. How can you easily sell more ads? By having your algorithms favor posts and media that drives engagement over the mundane truth. Which is sad.
I guess, in sum, this is a bad but predictable move by Facebook. It does nothing beneficial for its users while offering a fig leaf towards gaming engagement via algorithms.
Addendum:
Here's a good interview with three Mueller team members on the investigation writ large: https://www.npr.org/2024/09/24/g-s1-24189/mueller-investigator-says-russia-interfered-in-2016-and-in-the-2024-election-too
Relevant bits:
ZEBLEY: Well, in that instance, we said we did not find sufficient evidence of a conspiracy to charge a conspiracy. Collusion is not really a criminal violation. So we applied that term to mean conspiracy. And we ultimately said, look, we didn't find sufficient evidence to say there was a conspiracy. But we did say that the Russians interfered in the election with the intention of favoring Donald Trump because they perceived there'd be a benefit from a Trump presidency, and they interfered in a way to harm the other candidate, Hillary Clinton. And we also found evidence that the Trump campaign perceived it would benefit from some of the Russian conduct. For instance, they perceived they would benefit from the releases by Wikileaks, and at the time, Wikileaks was releasing information that the Russians had stolen from the Clinton campaign.
GROSS: There was also the infamous Trump Tower meeting where a Russian proposed to Don Jr. that this Russian had damaging information about Hillary Clinton that the Trump campaign could use and offered to have a meeting to hear about it. And Don Jr. agreed to do it. And so Don Jr., Eric Trump, and Paul Manafort, who was the campaign manager at the time, met with a couple of Russians. As it turns out, the Russians had their own agenda, but the fact that the Trump sons and campaign manager were willing to have that meeting, in terms of information, what did that tell you?
(made some edits in my post for clarity)
These were just a few examples from a much longer piece on this topic summarized - I figured someone would comment on one here, which is great bc it proves the point - that a community (of smart people like you, whose ideas I respect & appreciate) can add input to things, which just happened. I would just rather have this effort done in a decentralized fashion by many citizens than a few people (who frequently have specific political agendas).
I hesitate to take Meta at face value on anything, but I agree that the fact-checking industry has been distorted and overextended from the very specific (and valuable) role it's long had in journalism and other publishing. And it's quite possible the downstream effects of this reputation harm are worse than if there had never been high-profile fact-checking operations.
Obviously, the idea of fact-checking satirical content is ridiculous and pointless, as someone with a vested interest in this! (https://substack.com/@jamescooperdasilva/note/c-84443777)
I'd also argue that Biden's cognitive health (or Trump's in his 1st term) is an area where fact-checking isn't well-suited. It's one thing to accurately convey official doctor's records, etc., as is always done with presidents, but the armchair diagnoses over the past 8 years have been guesses at best, not factual analyses.
As for the inflation and COVID origins situations, those illustrate how fact-checking is used prematurely for ongoing situations. It's one thing to say, for instance, "So-and-so claimed China did COVID, of which there's not proof. Here is what we know about the possibilities." But too many of these went further, as you note, as if there wasn't more to learn.
Great post. After nearly 30 years on the internet the best censorship is the unfollow, mute, and block buttons.
Exactly right, you already personally control this
Great post Adam. I'm a massive fan of Community Notes on X. IMO it is the single most important feature created by any social media company since the invention of the newsfeed. CN should be on the front page of ALL media outlets (social or publishing).
An unappreciated reason why it succeeds is that it's anonymous. A CN must stand on its own merits, independent of who wrote it. Likewise anonymity means that there's no social status to be gained within the community of Note writers. Compare this to Wikipedia where mods entire identities are created around the social structure of editors.
And a final thought - why is it that we use secret ballots at elections? To avoid preference falsification.
Community Notes for the win.
Well said.
The closer our institutions and platforms are to a meritocracy the better. Community notes and crowdsourcing are not 100% accurate or precise, but more times than not they correct narratives that are misleading or flat out false. No one is immune to propaganda, but this mechanism is a good counter to it or at the very least provides good counter points to prevailing arguments.
That's how nations are built, through storytelling.
In France, we have this famous expression "le roman national" literally 'the national novel'. Pure fiction
Once again, you're unfortunately missing the point. We have numerous studies that show fact checking is not biased, but that conservatives share significantly more misinformation than others. That is why they are fact checked more often! Pursuing political balance in fact checking actually makes it more political, not less.
And the claim that fact checkers in Texas are some how less biased than California?? That's absurd.
Community notes are a good step forward, but they are NOT a replacement for fact checking. While fact checkers can get it wrong, and community notes can get it right, that is not evidence that fact checkers are always wrong or that community notes are right more often!
As long as social media companies have algorithms that choose to promote some content, they are responsible for spreading misinformation. Calling that responsibility "censorship" is not a productive framing.
Thanks for the comments Sean I always appreciate your perspective. If you dig enough into politicians you'll see both sides lie, more resources are simply allocated to one side here so there is a lot of selection bias. If you get a large enough sample of politicians this has gotta be evenly distributed. Unsure how you'd test but if you listen to Congressional hearings on CSPAN etc unedited by TV news you get a sense of how there's reasonable dems and cons, it actually makes me way more optimistic for our institutions than cable news etc.
The culture in California is *definitely* super biased, I've lived in 4 major cities in America >2 years and can see that. Moving the teams to Texas was symbolic sure, but honestly I bet you get more centrism here. Anyway separately Zuck stated they're gonna focus trust and safety on other things. Because of this they will have *more* resources to tackle actually bad and illegal actors, predators, spammers and stalkers. This is what they should be doing, we all agree that stuff is bad.
Glad you agree community notes are cool. The algo discussion is an entirely other team than the in-house and outsourced censorship folk ofc, so a very different discussion. I agree with you here too those algos particularly on FB should bias to content from friends vs politics or viral slop. IDK why they do that.
This isn't a matter of opinions, we have real scientific studies that show conservatives spread more misinformation. Here is one, I suggest you read a few of them: https://www.nature.com/articles/s41586-024-07942-8
As for Texas being less biased than California, it really is absurd. Both states are enormous and have a multitude of people with many different beliefs. The idea that geographic location determines your bias is insane and evidence of opinions over facts. Your personal network is not representative of an entire state!
This isn't just a topic where opinions can differ, that's part of the problem. There are real facts here and you can't disagree on facts.
Texas has a more balanced mix of urban liberal areas and rural conservative areas, while California is overwhelmingly Democratic. I am not considering my personal network I'm looking at this at the macro level and fairly dispassionately. You're right though this shouldn't determine N=1 person's bias. It's symbolic anyway, FB just needs to hire better. You go to work to make revenue for the company, not participate in activism.
There is bias in these studies btw, how do we define misinformation? There's left leaning people arguing against 2+2 = 4 in academia and other areas. No one is without fault. Also using Twitter here presents additional selection bias, and a lot of what is in that link is outright spam. Regardless none of it changes that I don't think tech companies should be employing ideologically-driven fact checkers.
I believe you that they are your opinions, but we have facts on this. We can't dismiss studies because you think there is *hand waving* bias. That's exactly the problem that fact checkers need to address.
Texas is not less biased than California. Based on voting records, Texas is at least as conservative as California is liberal but even that is unreliable.
I appreciate we can have a friendly discussion about this, but it's evidence of why fact checking is necessary. There is truth here and opinions, and normal folks don't have the time to separate the two.
Appreciate it too. And look on opinion pieces those are just that, they're labeled as such or on an opinion blog and shouldn't need community notes so much, they're presenting ideas and narrative and in good cases have factual proof points/links too. Specific breaking news stories or comments and promises from elected officials *should* be held to account. We agree on this! A new experiment is about to unfold so we'll see what happens. I remain optimistic.
Sean, the main issue that I’ve had is that, on important issues, fact checking and censorship is one sided.
For instance, on Covid 19 vaccinations, I don’t recall anyone being fact checked in 2021 for saying that vaccines were “effective” at preventing transmission, but plenty were banned for (correctly) saying the opposite. The impact was that many people then self censored.
On climate change, I don’t see any media being fact checked for (incorrectly) saying that hurricanes have grown more frequent and/or more intense in the Atlantic basin. On the other hand, plenty have been censored for correctly presenting the scientific evidence. See Roger Pielke Jr’s post from yesterday.
Would you argue fact checking has been balanced on these two topics? If not, are there meaningful policy discussions in which fact checking is balanced?
You seem to have some anecdotes about fact checking like Adam does, and anecdotes are not data. I'm not going to debate what you recall or experienced, which sounds different from me.
What I will say is that there is evidence that hurricanes are more intense due to climate change, so your perspective might need to be updated. Here's a good example: https://iopscience.iop.org/article/10.1088/2752-5295/ad8d02
Our discussion shows the limitations of fact checking and highlights the importance of robust, uncensored discussions.
About hurricanes — I would guess the article you link bases its claim on data collected from about 1980. There’s a valid reason to do that (we didn’t have accurate satellite imagery before). However the weakness is that hurricane activity shows strong decadal patterns, and 1980 was a low in hurricane activity. I’m basing my claim on NOAA and the IPCC AR6 WG1, which take into account data starting from the early 1900’s.
We can spend days discussing this with intellectual honesty and still end up disagreeing. In other words, this is an area where fact checking is not useful, and actually prevents this kind of necessary discussion.
Fact checking is useful for trivial statements like “there are 53 states in the US” or “the iPhone was invented in 1453,” not much else.
No, I linked to a study from last year that focuses on hurricanes in the last decade. What this demonstrates is the need for fact checking because you are convinced of an incorrect conclusion and haven't revisited your priors. You cannot rely on community notes to remedy those things, as all they end up doing is reinforcing popular priors.
Generally speaking, the need for fact checking is that social media allows they misconceptions to spread and become priors for other people. Just like you dismissed the study I shared, people tend to avoid uncomfortable truths and community notes doesn't address that at all.
Sean, this is a worse study than I was giving it credit for. “Focuses on hurricanes in the last decade”: there is no way that one can draw any conclusion on the relationship between global warming and increased hurricane intensity over such a short amount of time (10 years). That would be the case for any climate phenomenon but even more so for hurricane activity which shows strong auto correlation.
There’s one way to avoid cherry picking studies of dubious quality and that is to rely on assessments such as the IPCC.
Honestly, if you aren't going to read the studies then I'm going to leave the conversation here. It doesn't just study the last decade, but it focuses on whether the past decade fits historical patterns. It's a good study, and I recommend doing a lot more reading on the topic before you form conclusions.
I wonder if this will actually apply to topics like criticism of Israel, which Meta was recently found to be suppressing. It’s hard for me to believe Zucc won’t carve out exceptions.
Community notes is great, and is one of the few improvement made to the platform. I don't trust Zuck, nor the motives behind this, but, as you state, as long as Meta is still able to (or willing to) remove genuinely harmful, dangerous content, then supplementing that with community moderation seems a sound decision.
It's disingenuous to say that community notes will work to monitor and counter disinformation and hate speech when it hasn't worked on X. By the time community notes can submit counterfactuals, it's too late. Misinformation has already gone viral. Besides, people simply take screenshots of the post before it's flagged and propagate it this way. It's easy to get around community notes. Community notes sound good, but in practice, it doesn't work.
Your analysis was remiss to note the political context of Zuckerberg's decision. I conclude that Zuckerberg's decision to remove fact-checking is politically expedient for him to do so. Consider the fact that last week, Meta’s global policy chief, Nick Clegg — a former British deputy prime minister who was chosen for his centrist bona fides — was replaced by Joel Kaplan, a longtime Republican operative who has acted for years as Mr. Zuckerberg’s liaison to the pro-Trump right. On Monday, Meta announced the appointment of three new board members, including Dana White, the chief executive of the Ultimate Fighting Championship and a close friend and political ally of Mr. Trump’s.
Zuckerberg promotes his decision as a return to free speech ideals and the need to avoid too much censorship. However, he fails to acknowledge what had happened on Meta’s sites when misinformation or foreign propaganda was permitted to propagate unchecked.
Unfettered hate speech has real-world consequences. Take, for example, what happened in Myanmar in 2017, where the Rohingya were killed, tortured, raped, and displaced in the thousands as part of the Myanmar security forces’ campaign of ethnic cleansing. In the months and years leading up to the atrocities, Facebook’s algorithms were intensifying a storm of hatred against the Rohingya, which contributed to real-world violence. How are Community Notes supposed to combat this?
For those who want a counter-argument to Adam's defense of Zuckerberg's decision, I suggest you read The Big Picture's substack entitled "Royally Zucked": https://shorturl.at/kuocl.
Just want to note that unfettered hate speech is not even allowed on Twitter/X, their team removes it if it violates their guidelines (https://help.x.com/en/rules-and-policies/hateful-conduct-policy). Very safe to assume Facebook will still do the same. It really appears to me they are simply allowing discussion around a bunch of topics some people don't like. They even noted they are doubling down on removing grifters, spammers, predators etc which these resources should be devoted to.
Meta changed their hate speech guidelines for the worse. See Wired's article https://www.wired.com/story/meta-immigration-gender-policies-change/. One example: Under the old policy, users were prohibited from referring to women as "household objects or property.” That part of the former policy is now crossed out, so we can assume that the objectification of women is now fair game.
Of course, I'm not going to defend that. But I wonder to what extent certain policies like this would change much, were people really chomping at the bit to say certain things like this? Anyone who would harass you before you likely already blocked or muted.
It's a $1.5T company and I don't think they make these decisions lately. The tech press foams at the mouth to write negative stories here about any policy decisions that happen and will bias to be negative before things even occur. Of course, I personally bias to free speech (not to harass people, just to not shrink the spectrum of our ideas). That's probably why we post more here and not there, and with this there is very little bad behavior. Appreciate the thoughts.
They have to cater to their user base
"People don't understand me because they're wrong" strikes again
Pretty easy to be on the right side of this one, if you're on the side of the censorship-industrial complex at least consider if you are with the bad guys. This post isn't political. Also, if you can't understand why an election was lost, well, there's probably no helping you.
There's nothing more political than this post
Also pretty easy to claim you're on the "free speech" side when you're from the dominant class
The fact that all the misinformation researchers and fact checkers lean one way shows why we should wish the community to do this and not individuals. So that becomes inherently democratic and truth-seeking which is what America is.
Is it possible that one side is a lot more likely to spread misinformation though? While I think you make some good points on the flaws of a fact checking approach to Social media, I think you're completely ignoring the objective truth that there is a deluge of misinformation coming from one side. Yes there's misinformation from the left too, but volume and intensity wise it's not comparable.
Yeah of course, thus we want the community to do this, who is comprised of sufficient voices to be able to get a balanced view of things.
Truth-seeking? Nice story, bro. USA was built on lies. Hollywood spread fake news long before the eX-app. Remember Iraq 'weapons of mass destruction'?
BTW why referring to USA (that are not 'America')? I thought you were talking about social media.
This looks like more a Biased Take than a Hot one.
Nihilism to think USA was built on lies, honestly. I'm sorry you feel this way.