What I Saw in Elon Musk’s Truth Army
Reading Time: 14 minutesThere’s a Hidden Part of Twitter Where You Can See All the Failed Attempts to Fact-Check Elon Musk, Twitter’s ‘Community Notes’ volunteers are supposed to make the platform ‘the most accurate source of information about the world.’ I became one., Twitter’s
Two months ago, Elon Musk sat down for an interview at the Tesla Gigafactory in Texas, where CNBC’s David Faber asked him about free speech, among other topics. Faber wanted to know how the Twitter CEO’s philosophy on speech squared with his professed desire for ‘Twitter to be as truthful as possible, ‘the most accurate source of information about the world.’ What does that mean for how you police on the platform?’ Musk took the opportunity to praise one of his favorite Twitter tools: Community Notes.
‘You can think of Community Notes as, like, an error correction on information in the network,’ Musk told Faber. ‘If somebody knows that they’re gonna get Noted, they’re less likely to say something that is false, because it’s embarrassing to get Community Noted. And that applies even to advertisers.’
If you’re still a frequent Twitter user, you’ve surely noticed these ‘Community Notes’ pop up on the platform. They appear under a given tweet from a prominent account—blue check or no blue check—to add more information and linked citations, under the heading ‘Readers added context they thought people might want to know.’ In the days after his May 16 interview, you may have caught some of these notes under Musk’s own tweets—for example, when he quote-tweeted a Fox News graphic that claimed that a pro-LGBTQ+, anti-bullying education organization is ‘encourag[ing] secret gender changes among children in schools.’ Musk had written, ‘Is this true, @Target?’ Community Notes volunteers corrected him, adding appendices to his tweet to clarify that ‘Target has donated to [the organization] for more than a decade’ and that it ‘provides curriculum support in many subjects, including mathematics’—Notes that also appeared under the original Fox post.
Other Musk tweets about Ron DeSantis’ fundraising and tech journalist Taylor Lorenz have also been Noted. In each instance, the user-written corrections have shown that the CEO was spreading falsehoods, by citing major newspapers and official organizations’ webpages. That, needless to say, has not embarrassed Musk himself into posting less, nor has it made him less likely to spread lies.
I’ve been tooling around with Community Notes for months, hoping to understand its impact on Twitter’s fragile information ecosystem. One fun thing about being an active contributor is that you have the ability to see pending Notes, including the myriad ones submitted under Musk tweets over the past few months. These Notes haven’t gone public yet, since they haven’t reached a threshold requiring approval from a certain amount of users from across the ideological spectrum. These tweets include his controversial post comparing billionaire George Soros with X-Men villain Magneto; a screenshot he posted of a graphic that misattributes a quote from a neo-Nazi figure to the Enlightenment philosopher Voltaire; and, ironically enough, his quote tweet of a user’s claim that ‘CNN seems to hate that Twitter community notes works much better in a free speech environment than anything they have ever done.’ Participants are granted the option to ‘Rate proposed Community Notes.’ With a click, they can see what other users have suggested, whether it’s offering more information on a particular tweet or explaining why they think added context isn’t needed. (An acronym often deployed in these discussions is ‘NNN,’ which stands for ‘No Note Needed.’) Should any of these suggestions be rated approvingly by enough members of the Community Notes community of varying political leanings, that Note will at last be affixed to the tweet for all tweeters to see. If they don’t get those votes, the Notes will be stuck in this ‘Help Rate’ limbo—or, worse, be deemed ‘Not Helpful’ and obscured from even Community Noters’ view. Notes can be sober and considered, shitpost-y and troll-y, false or bigoted—an aptly chaotic microcosm of the new Twitter world.
Musk’s Twitter tenure has seen the network devolve into a far-right-friendly, spam-ridden, hate-speech-saturated cesspool whose most outspoken users include transphobic tech bros and white nationalist pundits. Community Notes hasn’t (fully) followed in that path, yet. The tool has provided needed checks on prominent users of all ideological persuasions, extending from Occupy Democrats’ founder to British conservatives. Even Musk buddy David Sacks couldn’t preempt a public Community Note on his June 28 tweet about ‘American troops … in Ukraine‘ by offering media links for those who ‘want to fact-check this.’ Funny enough, the Note drew on those very sources to clarify that ‘1) US pulled out its military instructors PRIOR to war beginning in Feb 2022, and 2) military personnel that do remain are embassy security.’ Though he seems to appreciate Community Notes overall, Sacks was clearly unhappy about that pushback and later called out the Ukraine Note as ‘bogus’; there are a few proposed, nonpublic Notes suggested on that tweet explaining why the original Note was not, in fact, bogus.
Still, Community Notes is far from sufficient in checking the rampant spam and disinformation that flocks freely across the platform, already a serious issue in pre-Musk times and far worse now. As the Associated Press reported, even though Musk told CNBC that his system would correct tweets that push conspiracies about the 2020 election, those posts spread unabated across the platform in the week following Donald Trump’s CNN town hall, the most shared of which included no public Community Notes. In general, corrections with a political valence have trouble getting enough votes from an ideologically diverse group on Community Notes—even in cases in which a political or media figure has very clearly lied.
This pattern was evident just last month, with a large portion of inflammatory, misleading tweets regarding Trump’s federal indictment—originating from influential figures like Kari Lake, Steve Bannon, and Charlie Kirk, among others—going thoroughly unchecked. The few public Notes on the subject were arbitrary and unhelpful, explaining to Matt Schlapp the real reason Trump didn’t get a mug shot or informing Sen. Marco Rubio that the Department of Justice has made no sentencing recommendations for Trump. Even more weirdly, as Community Notes members can spot, suggestions have been made to append Notes to claims that President Joe Biden is ‘weaponizing’ the DOJ and that Rep. Adam Schiff ‘knowingly lied about the Russia Collusion Hoax.’ Unrelated to the indictment, other public Notes from the indictment period include a limp correction to Ian Miles Cheong’s lie that ‘Pepsi has released a new ad that makes America and the military trans’ (the correction: ‘This is an advertisement by SodaStream from 2021’), and a reminder on Nancy Pelosi’s remembrance of the 2016 Pulse nightclub shooting that the killer did not commit an ‘act of hate.’
Far more troublingly, Musk stans and far-right influencers sometimes scapegoated and weaponized Community Notes for their own personal gain, regardless of accuracy. The Twitter account for the pro–gold standard Wall Street Silver subreddit accused Community Notes of being overrun by ‘volunteer leftists,’ while the right-wing activist Andy Ngo accused left-wing users of utilizing the block function ‘to circumvent @CommunityNotes,’ leading Musk to suggest that blocking be done away with entirely. Other examples show right-wingers and Musk fans spearheading and cheerleading the addition of inaccurate Community Notes to multiple tweets by journalists like Judd Legum and Glenn Kessler (the Washington Post’s Fact Checker columnist—who, to be fair, does not have a spotless record).
Not that this has always played to Musk’s favor: On occasions when he’s tagged the Community Notes account to gauge the accuracy of, say, Robert F. Kennedy Jr. calling Anthony Fauci a bioweapons manufacturer or a ‘citizen journalist’s’ declaration that the FDA all but admitted that ivermectin treats COVID-19, no public Notes appeared for either claim; Community Notes members did submit multiple entries both supporting and detracting from the claims, although no consensus was reached. (To be clear, both of those tweets were very, very wrong.) Musk tried to summon Community Notes to ‘check’ himself on a misleading tweet about gender-affirming care for trans youth, but the site did not surface even one of the several proposed Notes that countered him. Even Musk fans don’t always love the feature, claiming that it’s too biased in favor of leftists, Big Pharma, or both. Sometimes, those most aggrieved with Musk, like the tweeters upset with the ‘rate limits‘ he suddenly imposed earlier this month, will propose Notes specifically meant to mock and insult him.
Nevertheless, Elon Musk continues to tout Community Notes as an indelible, indispensable function for ensuring Twitter’s future, attempting to reassure users that when he tweets out a false Federalist article about the 2020 election or when he platforms Tucker Carlson—a pundit who’s gotten in legal trouble multiple times for his on-air lies—Community Notes will correct everything without bias or favor. Forget that none of the many submitted Notes suggested for Tucker on Twitter‘s multiple broadcasts has been deemed ‘Helpful’ enough for public consumption. And forget, also, that these very Notes and their discussions can devolve into outright juvenilia.
Community Notes was built by Twitter before Musk’s takeover, and some of its problems predate his arrival. But it has also changed under Musk to suit his mission, a process I observed up close. If even his own lies are allowed to stand, can Twitter really become ‘the most accurate source of information about the world’?
In January 2021, mere weeks after Twitter was forced to grapple with its role in fueling the Jan. 6 Capitol insurrection, the company’s vice president of product, Keith Coleman, announced a new ‘community-based approach to misinformation‘ on the Twitter Blog, to be known as Birdwatch. Kicking off its pilot program in the United States, Coleman wrote that ‘Birdwatch allows people to identify information in Tweets they believe is misleading and write notes that provide informative context.’ To start off, ‘notes will only be visible on a separate Birdwatch site,’ on which anonymous ‘pilot participants can also rate the helpfulness of notes added by other contributors.’ Eventually, however, ‘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 addition to public access offered to the coding of and ‘data contributed to’ Birdwatch. The company would encourage users to sign up for the pilot while seeking feedback and advice from academic ‘experts in a variety of disciplines.’ Individuals would also develop a ‘reputation score’ based on their record in grading Birdwatch notes, with new users starting at 0 and getting closer to 5—the level at which they could write notes all their own—by consistently weighing in on Notes that are deemed ‘Helpful’ or ‘Not Helpful’ by consensus.
A couple of weeks after the announcement, Poynter published detailed critiques of Birdwatch based on analysis conducted by reporter Alex Mahadevan, who participated in the pilot and claimed to have ‘analyzed more than 2,600 notes made by Birdwatchers and reviewed 8,200 ratings.’ At that point, Mahadevan wrote, the pilot had about 1,000 users—and many times more issues. He pointed out multiple tweets ranked ‘not misleading’ by these participants that trafficked in election denial, while tweets that highlighted well-documented connections between Donald Trump’s rhetoric and the Jan. 6 violence were deemed ‘Harmful.’ Further, a single-digit percentage of the 2,600 Notes studied were deemed ‘Helpful,’ and none of the top 10 most active Birdwatchers were fact-checkers, journalists, or people otherwise trained in snuffing out and countering digital misinformation. Plus, the majority of Notes did not cite a single source outside of Twitter. In response, the Birdwatch team tweaked its algorithm to raise the threshold for a Note to be considered ‘Helpful,’ including the need to cite a quality source.
That June, as TechCrunch reported, Twitter shifted these Birdwatch ‘fact-checks’ from the designated website to Twitter itself, although they were visible only to pilot participants. Notes that reached a ‘Helpful’ consensus would stay visible, while others that were not deemed thus would disappear from view. By November, Twitter had established a private Discord server for Birdwatch users to congregate and deliberate over their decisions and research, especially regarding prominent tweets from politicians, celebrities, nongovernmental organizations, and corporations, or random posts that achieved virality. An employed community manager from Twitter kept in dialogue with these folks.
To take it from those Birdwatch watchers, the program was bound for success: In March 2022, in a blog post announcing ‘a slew of improvements‘ to Birdwatch, Coleman stated that ‘in surveys with randomly sampled people on Twitter in the U.S.,’ a majority of users found ‘Helpful’ Birdwatch notes to be, well, helpful, including for ‘people from across the political spectrum’ (in this case, Birdwatchers who had consistently voted the opposite way on notes preceding the ‘Helpful’ one). And those polled ‘were 20-40% less likely to agree with the substance of a potentially misleading Tweet after reading a note about it, compared to those who saw a Tweet without a note,’ he wrote. That enthusiasm was somewhat belied by an April MIT study finding a strong partisan influence in how users were likely to gauge a note’s accuracy—e.g., that Democrats rated 83 percent of Democrat-written notes as ‘Helpful,’ while Republicans rated 87 percent of notes from their side the same way.
By September, Twitter decided to expand Birdwatch Note visibility outside its then-15,000 users to half of Twitter’s U.S. base, and encouraged even more users to sign up. New features included an algorithmic function to ban a user from writing notes if their paper trail was often proved ‘Not Helpful,’ and a boost to notes collectively rated ‘Helpful’ by users of differing political ideologies. Despite the company’s enthusiasm, issues remained: Internal documents made public by whistleblower Peiter ‘Mudge’ Zatko showed that Birdwatch had accepted an explicit QAnon conspiracist as a member, while failing to implement content-moderation measures recommended by its expert consultants.
On Oct. 6—just a couple of days after Elon Musk confirmed he’d go through with his purchase of Twitter, after fighting the deal most of the year—Coleman wrote that ‘Helpful’ Birdwatch notes would be visible to all U.S. users on tweets they were applied to, with additional written information and linked sources ready for public consumption and contextualizing. The network’s end-of-month sale to Musk’s ownership (during which, notably, Keith Coleman stayed on as other executives were sacked) would provide a real-time example of Birdwatch’s limits: When Birdwatch notes appeared under a President Joe Biden tweet about corporate taxes and a White House tweet about Social Security, Musk stans celebrated these as directly attributable to their new king. This, of course, did not align with the actual Birdwatch timeline, but these posters were not people inclined to read (much less cite or trust) a Reuters fact-check explaining why this assumption was wrong.
In early November, Mahadevan penned a thread laying out how Birdwatch use ‘exploded’ after Musk’s takeover, with daily rates of Notes additions, ratings, unique engagements, and ‘Helpful’ consensuses all increasing, even as the overall number of active Birdwatch users shrank. (The Washington Post subsequently found that nearly half of the Notes ranked visibly ‘Helpful’ in the days following Musk’s purchase were written by only 25 users.) However, Mahadevan also noticed that a full 25 percent of public notes deemed ‘Helpful’ pre-Musk were disappeared into Birdwatch-exclusive ‘Needs More Ratings’ limbo—including one adding context to a Musk tweet about ‘appeas[ing]’ supposedly ungrateful activists who were pressuring advertisers to leave Twitter and, thus, ‘trying to destroy free speech in America.’ (It later reappeared in public view on that Musk tweet, thanks to a process Coleman referred to as ‘entirely organic’ and ‘powered by the people.’) In response to Mahadevan’s thread, Musk tweeted that Birdwatch would soon be renamed ‘Community Notes’ and had ‘incredible potential for improving information accuracy on Twitter!’
Community Notes’ value to Musk wasn’t just in its potential utility for inviting advertisers concerned about (in)accurate information on Twitter, but in its ability to cut costs: an entirely volunteer-run army of amateur fact-checkers to replace the paid content moderators to whom he’d shown the door. The Chief Twit’s controversial moves here didn’t sit well with many longtime Birdwatchers, who complained to the Post that ‘we don’t have a moderator anymore.’ A college student involved with the Birdwatch Discord from jump mentioned that ‘it feels like the entirety of the Twitter misinformation system is now reliant on Birdwatch. Birdwatch is important but shouldn’t be the adjudicator of facts.’
It was shortly after the November midterms that I was approved to participate in the rebranded Community Notes. I had applied on a whim when a prompt suggested that I try out Community Notes, with no expectation that I would be granted access. Yet there I was, a newly anointed fact-checker for Musk’s brave new Twitter.
On any given day, to access the Community Notes portal, I have to click a button illustrated with three user-outline icons bunched together on my left-side Twitter sidebar. I’m then taken to a feed that resembles Twitter’s homepage, except that it lists obscured and public notes arranged under three different tabs: ‘Needs your help,’ which displays ‘notes chosen for you’ on tweets that ‘need a more diverse range of feedback’ and where my ‘point of view could help decide if they’re helpful’; ‘New,’ which showcases ‘the most recently written notes’ for which we can provide feedback; and ‘Rated helpful,’ which shows Notes made available to public view that recently ‘have been rated helpful by contributors of multiple perspectives,’ and that I can further rate as ‘Helpful,’ ‘Somewhat Helpful,’ or ‘Not Helpful.’ (I can also see other nonpublic Notes that were submitted for that given tweet, regardless of whether they were designated ‘Helpful’ or not.) This portal also includes a component declaring the ‘Community Notes values’: ‘Contribute to build understanding,’ ‘Act in good faith,’ and ‘Be helpful, even to those who disagree.’ A page from the official Community Notes Guide spells these values out: It’s ‘not a place for quick dunks, personal opinions, or insults’ where people should ‘game the system’ and use ‘hateful, derogatory, or inflammatory language.’ Rather it’s an arena where users should ‘write clear, evidence-based notes to help everyone, even skeptics, better understand a Tweet and why it might be misleading,’ where the chosen ones should ‘genuinely and constructively contribute … in a way that even those who may disagree with you might find helpful and respectful.’ (The FAQ clarifies that ‘all contributors and all Community Notes contributions are subject to the Twitter Rules’ and that if you’re a tweet author who’s received a disagreeable public note, ‘you can request additional review.’)
The alias I chose for myself as a Community Notetaker is ‘Ecstatic Tree Robin.’ (The feature autogenerated my name choices.) As of this week, I have a Rating Impact of 13, thanks to 11 of my ratings that ‘helped a note earn the status of Helpful,’ along with two ratings that ‘helped a note reach the status of Not Helpful.’ (I have not made any ‘Ratings of helpful or not helpful on notes that end up with the opposite status,’ even though the program explains, about ‘contrasting ratings,’ that ‘everyone gets some of these!’ Well, not me, suckers—I’m just built different.)
As of this writing, I do have 125 pending ratings on ‘notes that don’t currently have a status of Helpful or Not Helpful,’ along with 24 other ratings ‘made after a status was reached,’ without any additional effect on those Notes as of yet. I cannot easily look up what those were. However, I can customize the types of Notes I receive on my feed through the settings: the language they’re in, and the frequency with which I receive alerts as to Notes that need additional ratings.
I see Community Notes not only on these special feeds, but out in the wild, in my everyday Twitter experience. Made available to my account are horizontal options listed below tweets that have received notes, which I can see and rate should I click. There might be just one note submitted to a tweet, or several; these are grouped into categories depending on whether they’re ‘suggesting context to be shown with the Tweet’ or ‘explaining why context isn’t needed.’ I can also get notifications both as to whether a Note I weighed in on was deemed ‘Helpful,’ or whether there’s a pending Note that could use my insight.
Since I have a Rating Impact that’s greater than 4—a threshold I breached just last month—I’m also allowed to propose my own written Notes. I have written only four so far: one on a conspiratorial June 19 tweet speculating on the cause of Houston rapper Big Pokey’s recent death; one on an implication from Chaya Raichik, creator of the infamous ‘Libs of TikTok’ account, that an increase of nonbinary students in New Jersey schools represents ‘a social contagion’; the third on a tweet by a Twitter Blue–subscribing Indian techie who claimed that Jack Dorsey’s statements on the Indian government’s Twitter censorship demands have ‘been busted many times’; the fourth on a now-unavailable lie from a British user who tried to spin a report on a United Kingdom Court of Appeal ruling regarding asylum-seekers as a ‘Supreme Court’ decision that would ‘convert the UK into an Islamic nation.’ As of now, I have a 0 Writing Impact score, as none of my notes has earned ratings from ideologically disparate Noters. The most engagement I have received from my fellow fact-checkers, it seems, is a countering Note on the Raichik tweet telling me that ‘The tweet is expressing an opinion. Rebuttals belong in the replies not community notes.’ Such is life.
On June 12, kicking off her second week as CEO of X Corp./Twitter, Linda Yaccarino wrote a 10-tweet thread outlining why she’d left NBC’s advertising department to take on this new gig, and what she hopes to accomplish in the network’s ‘Twitter 2.0’ era.
‘The global town square needs transformation—to drive civilization forward through the unfiltered exchange of information and open dialogue about the things that matter most to us,’ Yaccarino declared. ‘Twitter is on a mission to become the world’s most accurate real-time information source and a global town square for communication. That’s not an empty promise. That’s OUR reality.’
Much of this echoed what Musk pledged upon taking over the network in late October—that Twitter should be a ‘town square‘ devoted to ‘free speech,’ that the embattled social network should become unmatched in its accuracy, that he would run Twitter in an ‘open’ and transparent manner. We’ve … seen how that’s turned out. Yet the network continues to double down on Community Notes—last week, it announced that its test program in Japan was found to be ‘helpful’ and that Japanese Notes will be open to the global public.
There have been signs Yaccarino could deviate from Musk’s ‘extremely hardcore’ style. For one, she first wrote the thread as a memo to the company’s 1,500 employees, even as mass internal communication has become rare in the Musk era. Plus, as the Information’s Erin Woo reported, Yaccarino’s first hire forwarded that very email out to tech reporters, marking ‘Twitter’s first official external comms since [Musk’s] takeover‘ (and a relief to those of us bored with the site’s automatic poop-emoji email responses). A month before she was appointed CEO, Yaccarino celebrated a Twitter Safety announcement on adding ‘labels on some Tweets identified as potentially violating our rules around Hateful Conduct letting you know that we’ve limited their visibility.’ (It should be noted that Twitter, despite massive staff reductions, still has employees working on user safety, although some workers told the BBC in March that ‘the company is no longer able to protect users from trolling, state-coordinated disinformation and child sexual exploitation.’)
As my colleague Heather Tal Murphy and I have noted, Yaccarino has some ideological overlap with Musk’s right-wing turn, and has been supportive of Twitter’s efforts to center figures like Tucker Carlson, Robert F. Kennedy Jr., and Ron DeSantis, all well known for spreading misinformation about vaccines, immigration, and climate change, among other topics. Further, Yaccarino’s veneration of the ‘global town square’ implies Twitter will retain its looser controls around what users can post, leaving Community Notes’ imperfect volunteers as a primary cleanup crew.
I ain’t sticking around to help. To avoid any future conflicts of interest, I hereby relinquish my Community Notes/birdwatching duties. To Yaccarino, Musk, and my now-former Community Notes comrades: Best of luck!
Reference: https://slate.com/technology/2023/07/twitter-community-notes-elon-musk-fact-checking.html
Ref: slate
MediaDownloader.net -> Free Online Video Downloader, Download Any Video From YouTube, VK, Vimeo, Twitter, Twitch, Tumblr, Tiktok, Telegram, TED, Streamable, Soundcloud, Snapchat, Share, Rumble, Reddit, PuhuTV, Pinterest, Periscope, Ok.ru, MxTakatak, Mixcloud, Mashable, LinkedIn, Likee, Kwai, Izlesene, Instagram, Imgur, IMDB, Ifunny, Gaana, Flickr, Febspot, Facebook, ESPN, Douyin, Dailymotion, Buzzfeed, BluTV, Blogger, Bitchute, Bilibili, Bandcamp, Akıllı, 9GAG