zlacker

>>danso+(OP)
Twitter is well within the rights to do this, but I have seen tweets from blue check marks essentially calling for violence and Twitter didn't remove them. So, does that mean Twitter actually -supports- those view points now? If Twitter is going to police people, it needs to be across the board. Otherwise it's just a weird censorship that is targeting one person and can easily be seen as political.

Everyone is applauding this because they hate Trump, but take a step back and see the bigger picture. This could backfire in serious ways, and it plays to Trump's base's narrative that the mainstream media and tech giants are colluding to silence conservatives (and maybe there could even be some truth to that.) I know the Valley is an echo chamber, so obviously no one is going to ever realize this.

>>partia+uY
Erm, what? This is just not true, and is a false dichotomy. Moderation is hard. Always has been. Stuff will slip through the cracks.

POTUS has the most popular (and currently most controversial - note, that's _controversial_ not _extreme_ or some other morph) so it's easy to see why Twitter are on top of it. Other blue-checked accounts, whilst more "important" than unverified, just simply don't compare to the importance and prevalance of POTUS' account.

>>Jestar+xc1
If most of the mistakes happen in one direction, then I would argue that there's some other mechanism at work than just "mistakes".

Update: data https://quillette.com/2019/02/12/it-isnt-your-imagination-tw...

Update: admission https://www.vox.com/2018/9/14/17857622/twitter-liberal-emplo...

>>efitz+Bk1
I’m not convinced by the arguments from your first link. As stated by the article itself, a difference in the number of left-leaning vs right-leaning bans does not prove the standards for censorship are different depending on what side of the political spectrum you fall on. It could be that conservative content violates rules more frequently than liberal or centrist content.

It goes on to say this can’t be possible because it would mean that conservative content would have to violate rules at 4x the rate of others, and that statistically its highly improbable. Why? It’s a known problem that Twitter has a lot of accounts that are fake accounts from bad actors trying to sow discord in the US political system, and those tend to be right leaning. Didn’t Twitter relatively recently do a purge of a large number of accounts that were deemed fake? That could easily skew the numbers, especially because those accounts tend to engage in the kind of rhetoric that gets you banned.

And then the article points to cases where liberal leaning content doesn’t get banned even though it should. I can also find cases where conservative content violates the rules yet it didn’t face consequences, most prominently the president’s account. It’s not just liberals who get a free pass, so I’m not sure what that proves.

Is it possible there is a bias in how Twitter sensors content? Sure. But that article makes it sound like they have a data driven, mathematically rigorous proof that it’s true, and I don’t think they meet that mark.