4 ms·

Twitter actually has some of the best ad oversight, almost entirely by accident too. Community Notes (formerly Birdwatch) lets users flag tweets that are misle

Twitter actually has some of the best ad oversight, almost entirely by accident too. Community Notes (formerly Birdwatch) lets users flag tweets that are misleading. Since Ads on Twitter are just normal tweets, they can be fact checked. Twitter's average ad quality has plummeted recently, but most of the scam ads I see have a massive user-added disclaimer outlining how it's a scam. I wonder how this affects the click through rate for these ad placements?

IG_Semmelweiss 3y ago
I've been thinking about this for a while. I've seen many of these community notes - how exactly do these work ? Is a community note borne out of "most upvoted" replies to a particular tweet ? If so, how does it go from a simple upvoted reply, to a "community note" ?
- voxic11 3y ago
  Basically if you are in good standing on the platform you can apply to be a contributor and if accepted you can submit and rate notes. The note which actually appears to everyone is selected out of the contributor submitted notes using some algorithm which is open source. > To find notes that are helpful to the broadest possible set of people, Community Notes takes into account not only how many contributors rated a note as helpful or unhelpful, but also whether people who rated it seem to come from different perspectives. https://communitynotes.twitter.com/guide/en/about/introduction https://communitynotes.twitter.com/guide/en/about/introducti...
- iteratethis 3y ago
  You need to sign up to become a community notes contributor. Then you can add notes, and if enough people from a diverse set of other contributors find it helpful, the community note will publicly appear. It's one of few things at Twitter that work shockingly well and reliable. I've never seen an obvious partisan community note. Other networks should embrace this.
  - cactusplant7374 3y ago
    > It's one of few things at Twitter that work shockingly well and reliable. Elon doesn't think so. I would expect it to change soon. https://twitter.com/elonmusk/status/1733882978781053383 https://twitter.com/elonmusk/status/1733882978781053383
    - ceroxylon 3y ago
      The way he was corrected by a community note on that statement... kinda hilarious.
      - cactusplant7374 3y ago
        I don't think he knows who this guy is or even cares. He's just pandering. And it's pretty low effort pandering at that.
    - yreg 3y ago
      Not really, Elon is quite proud of community notes in general. His claim that a particular note was gamed means nothing.
      - cactusplant7374 3y ago
        He thought polls were awesome too but then he changed it to paid subscribers only because of “manipulation,” I think? Dramatic changes will likely happen.
- TheCaptain4815 3y ago
  If twitter survives, community notes could usher in a golden era of truth on the internet imo. On a recent podcast w/ lex friedman, Elon explained how Community Notes works and it's pretty ingenious. In short, from what I understand, each community notes contributor is vector mapped based on their biases/contributions. Now, the only way for a community note to appear is for contributors who typically disagree with each other (based on their vector mapping) to both sign off on that note. I'd imagine in the future (or maybe it's enabled now), these contributors would be "power ranked" so a newbie contributor couldn't community note a member of congress or something. As Elon said, this will lead to hoardes of people leaving twitter because they cannot handle community notes.
- nomel 3y ago
  Paper explaining it in detail: https://github.com/twitter/communitynotes/blob/main/birdwatch_paper_2022_10_27.pdf https://github.com/twitter/communitynotes/blob/main/birdwatc... It's somewhat adversarial, with the idea that if people who have disagreed in the past agree on a note, then it's probably true.
iteratethis 3y ago
I think when an ad is fact-checked and fails that test, it should become hidden. Even better would be to block the advertiser from the network or even seize their ad spent.
- wongarsu 3y ago
  Not everyone agrees with every fact-check, even for objectively verifiable facts (which is the minority of all things claimed). Even Mythbusters have done things where I think they clearly reached the wrong conclusion. Having large consequences attached to the result of a fact check sounds like a recipe to alienate everyone. If you say something obviously untrue there are already laws against fraud and false advertisement. For anything with less burden of proof, a note below your ad that adds context seems adequate
- GaryNumanVevo 3y ago
  I think it's a great idea, for obvious reasons it will never be implemented.
- j-bos 3y ago
  Seems like best of both worlds for scam ads with notes to still appear. The scammers keep paying the platform, user's learn how to spot scams via the notes, and the platform makes money.
berkes 3y ago
> but most of the scam ads I see So it doesn't work. If it worked, you'd see no, or very little, scam ads.
- GaryNumanVevo 3y ago
  Yes, that's why I said accidentally. Community Notes wasn't intended to be a spam filter for ads. It's not able to get tweets/ads taken down, only place a large disclaimer below any fraudulent advertisements.
machdiamonds 3y ago
It's definitely not by accident, it would be very easy to turn off community notes for tweets flagged as ads.
mjcl 3y ago
I think the scammers are starting to get wise. I've seen over 50 different accounts running the same ad copy for some car solar defrosting gadget. There's just too many ads for any single one to get a published note and if one does they can delete the individual post.
mrtksn 3y ago
There's no day I go without a phishing attempt by ad on Twitter. The only time I've seen community note on ad is about some game that is apparently quite different from what it is on the ad. IIRC that game is notorious for this and has become a meme.
thrillgore 3y ago
It works until Elon has to step in.

IG_Semmelweiss 3y ago

I've been thinking about this for a while. I've seen many of these community notes - how exactly do these work ? Is a community note borne out of "most upvoted" replies to a particular tweet ? If so, how does it go from a simple upvoted reply, to a "community note" ?

voxic11 3y ago
Basically if you are in good standing on the platform you can apply to be a contributor and if accepted you can submit and rate notes. The note which actually appears to everyone is selected out of the contributor submitted notes using some algorithm which is open source. > To find notes that are helpful to the broadest possible set of people, Community Notes takes into account not only how many contributors rated a note as helpful or unhelpful, but also whether people who rated it seem to come from different perspectives. https://communitynotes.twitter.com/guide/en/about/introduction https://communitynotes.twitter.com/guide/en/about/introducti...
iteratethis 3y ago
You need to sign up to become a community notes contributor. Then you can add notes, and if enough people from a diverse set of other contributors find it helpful, the community note will publicly appear. It's one of few things at Twitter that work shockingly well and reliable. I've never seen an obvious partisan community note. Other networks should embrace this.
- cactusplant7374 3y ago
  > It's one of few things at Twitter that work shockingly well and reliable. Elon doesn't think so. I would expect it to change soon. https://twitter.com/elonmusk/status/1733882978781053383 https://twitter.com/elonmusk/status/1733882978781053383
  - ceroxylon 3y ago
    The way he was corrected by a community note on that statement... kinda hilarious.
    - cactusplant7374 3y ago
      I don't think he knows who this guy is or even cares. He's just pandering. And it's pretty low effort pandering at that.
  - yreg 3y ago
    Not really, Elon is quite proud of community notes in general. His claim that a particular note was gamed means nothing.
    - cactusplant7374 3y ago
      He thought polls were awesome too but then he changed it to paid subscribers only because of “manipulation,” I think? Dramatic changes will likely happen.
TheCaptain4815 3y ago
If twitter survives, community notes could usher in a golden era of truth on the internet imo. On a recent podcast w/ lex friedman, Elon explained how Community Notes works and it's pretty ingenious. In short, from what I understand, each community notes contributor is vector mapped based on their biases/contributions. Now, the only way for a community note to appear is for contributors who typically disagree with each other (based on their vector mapping) to both sign off on that note. I'd imagine in the future (or maybe it's enabled now), these contributors would be "power ranked" so a newbie contributor couldn't community note a member of congress or something. As Elon said, this will lead to hoardes of people leaving twitter because they cannot handle community notes.
nomel 3y ago
Paper explaining it in detail: https://github.com/twitter/communitynotes/blob/main/birdwatch_paper_2022_10_27.pdf https://github.com/twitter/communitynotes/blob/main/birdwatc... It's somewhat adversarial, with the idea that if people who have disagreed in the past agree on a note, then it's probably true.

iteratethis 3y ago

I think when an ad is fact-checked and fails that test, it should become hidden. Even better would be to block the advertiser from the network or even seize their ad spent.

wongarsu 3y ago
Not everyone agrees with every fact-check, even for objectively verifiable facts (which is the minority of all things claimed). Even Mythbusters have done things where I think they clearly reached the wrong conclusion. Having large consequences attached to the result of a fact check sounds like a recipe to alienate everyone. If you say something obviously untrue there are already laws against fraud and false advertisement. For anything with less burden of proof, a note below your ad that adds context seems adequate
GaryNumanVevo 3y ago
I think it's a great idea, for obvious reasons it will never be implemented.
j-bos 3y ago
Seems like best of both worlds for scam ads with notes to still appear. The scammers keep paying the platform, user's learn how to spot scams via the notes, and the platform makes money.

berkes 3y ago

> but most of the scam ads I see So it doesn't work. If it worked, you'd see no, or very little, scam ads.

GaryNumanVevo 3y ago
Yes, that's why I said accidentally. Community Notes wasn't intended to be a spam filter for ads. It's not able to get tweets/ads taken down, only place a large disclaimer below any fraudulent advertisements.

machdiamonds 3y ago

It's definitely not by accident, it would be very easy to turn off community notes for tweets flagged as ads.

mjcl 3y ago

I think the scammers are starting to get wise. I've seen over 50 different accounts running the same ad copy for some car solar defrosting gadget. There's just too many ads for any single one to get a published note and if one does they can delete the individual post.

mrtksn 3y ago

There's no day I go without a phishing attempt by ad on Twitter. The only time I've seen community note on ad is about some game that is apparently quite different from what it is on the ad. IIRC that game is notorious for this and has become a meme.

thrillgore 3y ago

It works until Elon has to step in.