I think I might have let out an audible groan in November when I saw Elon Musk proclaiming that his mission was to make Twitter “by far the most accurate source of information about the world”.
This is a man, after all, who has not in the past shown a particularly ardent commitment to the truth. In 2018, in a now infamous tweet, Musk said he had “funding secured” to take Tesla private (a deal that never materialised). And in 2020, he tweeted that children were “essentially immune” from Covid-19 (this turned out to be false).
Yet there is a feature of the social media platform that Musk seems genuinely excited about, and which is changing the game when it comes to internet-based fact-checking: Community Notes.
Community Notes is a feature that allows users to “add context” below other people’s tweets if they think they contain false or misleading information. Its origins predate Musk’s acquisition of Twitter — it was first piloted in the US in January 2021, though until November was called “Birdwatch” (a name that Musk said gave him “the creeps”) — but it has been expanded since the Tesla co-founder took the helm.
You can think of the feature as a bit like a crowdsourced misinformation-fighting system: a contributor can add a note when they feel a tweet is false or needing additional context, and other contributors are then asked to vote on whether this note is “helpful” or not. If enough contributors decide that it is, the note will appear underneath the original tweet (even in tweets that are embedded on other sites — a feature that was announced last week). If it’s not, other Twitter users will never see it.
Community Notes differs from traditional fact-checking — which I believe can be valuable but has too often been used as a political weapon — in several important ways.
First, and most obviously, the fact that it is done via consensus rather than by one person reduces both the likelihood and the influence of individual error or bias. Second, the feature does not actually call itself a “fact check”; rather, we are told that “readers added context”. This is an important difference — it’s an approach that treats facts as complex and contested and evolving, rather than imagining that they are always straightforward, undisputed and established. Finally, unlike not just regular fact-checking sites but also the rest of Twitter, Community Notes are totally anonymous, meaning there isn’t the normal incentive to virtue-signal or score points.
It’s not a perfect system. Twitter says that it’s only once enough contributors “from different points of view” rate a note that it earns the status of “helpful”. It says it assesses not by gathering information such as gender or political affiliation, but by looking at how people have rated notes in the past. This all sounds sensible, but it is difficult to ensure that the pool of contributors is diverse enough to start with. Just how different are these different points of view, and is the median one in the centre or is it some way off to the side?
Furthermore, how to ensure balance in terms of which tweets get context added to them? It’s unclear. Most of the tweets that I have seen getting “Community Noted” (yes, it’s already a verb) are from people on the left — like MSNBC host Mehdi Hasan last week, after a claim he made about rates of intraracial violence. Hasan responded to what was, in my view, fair and warranted added context by saying the Community Notes feature “has become another weapon of the right on Musk’s Twitter”.
This could end up being a self-fulfilling issue: if left-leaning users do not engage as much as others, Community Notes will become skewed towards the right. Molly White, a contributor who joined while it was still called Birdwatch, who is also a Wikipedia editor and who describes herself as a “leftist”, tells me that she has largely stopped contributing to the system since Musk’s takeover. “I was never particularly interested in providing labour for free for Twitter even under its previous ownership, but especially not now under Elon Musk,” she says.
But imperfect though it might be, Twitter is providing a model for a fairer, more transparent way to correct untruths and provide missing context on the internet. The company has even taken the quite radical step of allowing adverts to be Community Noted (a win for consumers, surely, though advertisers might not feel the same way).
Donald Trump is currently given 5/2 odds of winning next year’s US presidential election; being able to quickly correct false and misleading information before it spreads online is more important than ever. Musk might not always practise what he preaches. But that doesn’t mean his gospel is always wrong.