According to a story from Vice, Twitter has become more active in "shadow banning" Republicans--that is: reducing their electronic reach by, for example, not propagating tweets to other's feeds (when they ordinarily would be) or having their accounts show up in search results. This is a "shadow ban" because, in theory, the user does not know they have been actioned--they just go on their merry way (but with less interaction / impact).
Is this happening? Well, yes--according to a May 15th post twitter is using a number of tools--including algorithms to try to reduce "abuse reports." According to the Vice story, some prominent republicans have been excised from search lists while their Democratic counterparts have not. Why?
Asked why only conservative Republicans appear to be affected and not liberal Democrats, the spokesperson wrote: “I'd emphasize that our technology is based on account *behavior* not the content of Tweets.”That is: the R-tweets were, algorithmically, more trollish (or "negative," however Twitter defines that) than the D-tweets . . . according to 'science.' We don't know the specifics of the tests so we can't really be sure what the science is. We also don't really know when someone has been shadow-banned since they don't publicize it. A NOTE: Everyone thinks they've been shadow-banned.
On Facebook we see something similar: Facebook wants to fight fakenews--but they allow InfoWars to keep on truckin' despite claiming that the Sandy Hook shooting was a hoax (and other not-so-good but similar nonsense). Sure: some groups get banned. Some don't--but on the whole, Facebook's attempts are spotty and inconsistent at best--or cynically designed to clamp down on some bullshit--but not the especially monetizing bullshit. Maybe.
Could This Be True (That R's Are Worse?)
Twitter is, presumably, doing this in a smart way--that is, if Twitter does not have the very best people in the world working on this, at least they should have the second best. That said, we do not know precisely what they are aiming for. The Omnivore can make a few guesses though:
- Foreign influence. Twitter is likely alert and sensitive to the problem of foreign (Russian, other) influence. They want to avoid regulation and that's one major threat to it. If we believe the Hamilton 68 tracking (showing what bots associated with Russia are amplifying on Twitter) then it makes sense that Republican feeds would more-likely line up with foreign influencers (since according to Hamilton 68, Russia is mostly interested in amplifying Republican themes). This could asymmetrically hurt Republicans.
- Fake News. Facebook has more of a problem with fake news than Twitter does--but it's still top of mind. As the Republican news ecosystem is more infected (infested?) with Fake news than the Democratic one, it would make sense that if the algorithm is flagging fake news propagation as a negative that R's would get hit more.
- Conspiracy Theory. Again, R's lead here. If the algorithm is tracking theories like QAnon (or its little brother #pizzagate), the assassination of Seth Rich, and so on--again, the results could be asymmetric. This isn't to say that democrats don't have / tweet conspiracy theories--but probably not at the same magnitude and in the same amounts.
- Racism. Although Republicans will strongly protest, The Omnivore thinks the numbers are in and it's pretty clear that an account tweeting racial animus towards minorities is more likely to be R-leaning than not. Now, this can well go the other way (minorities tweeting hate-speech at white people)--or vice versa, SJWs tweeting #allMenSuck or whatever. If Twitter weighs these equally would it be a wash? The Omnivore doesn't know--but The Omnivore suspects that (a) Twitter would algorithmically privilege minority protection over majority and (b) might well privilege protecting women on its platform over men. Is that fair? Up to you--but ask your mom whether or not life is fair. She'll tell you.
This is all to say that just as there are real, significant differences between the genders, there may also be real, significant mathematical differences between the partisan tweeters. This should factor in to anyone's assessment.
What Does The Omnivore Think?
We have been treated to a slew of stories about how hard young Trumpers have it on the dating scene. The stories show the unfairness that these people have in finding a date as so many profiles say "if you're MAGA, get lost!"
In this case there is no algorithm at work--and no (specific) community standard in play--it is just the aggregate behavior of multiple users that creates a community dynamic where a Trumper is 'discriminated' against. A Trump-supporter might, for example, point out that a person who [ does something objectionable ] can find a date while they [ who are squeaky clean otherwise ] cannot.
If the "can't find a date" rule was enforced by Tindr, the discrimination would, The Omnivore says, appear "quite clearly" as unfair.
However, the fact is that on Tinder the "can't find a date" effect (which is analogous to Shadowbanning in our metaphor) is enforced by many individuals so it's harder to accuse the company of doing anything.
But, is it any less wrong? If Twitter takes control of their environment to limit disruption and negative interaction and has a reasonable methodology for it, is that good stewardship? The problem, of course, is that Twitter can't publish its algorithm because then people with a vested interest in circumventing it could get around it--but if Twitter has a view of what a healthy environment looks like--and GOP people are harming that, what are they to do? Throw up their hands and say "it hurts the environment--but, hey, it is fair?"
The Omnivore thinks this is a sticky situation--Twitter is a public forum for important speech. It is also a private company. Users may have moral cause to ask for transparency ("How can I follow the law if I cannot be told it?") but they don't have a legal cause for it--and the presence of people trying to undermine the rules makes keeping them secret a practical position for Twitter to take.
At this point, there aren't easy answers--but The Omnivore thinks that if you think you got shadowbanned there was probably a decent reason for it.