After President Donald Trump accused Twitter of “shadow banning” prominent Republicans, the company denied that it uses the practice, in which someone’s posts are made invisible or undiscoverable without them knowing.
In a blog post titled “Setting the record straight on shadow banning,” Vijaya Gadde and Kayvon Beykpour, Twitter’s legal and product leads, respectively, were blunt: “We do not shadow ban. You are always able to see the tweets from accounts you follow (although you may have to do more work to find them, like go directly to their profile). And we certainly don’t shadow ban based on political viewpoints or ideology.”
Twitter exists to serve the public conversation, enabling important discussions around the world to occur. Favoring one specific ideology or belief goes against everything we stand for. https://t.co/lioupj3aUF
— Vijaya Gadde (@vijaya) July 27, 2018
Gadde and Beykpour also addressed recent complaints, which gained more attention after a Vice article, that some accounts did not appear in auto-suggestions even when users searched for them by name. The two said the issue had been resolved and had affected “hundreds of thousands of accounts,” not just those representing certain ideologies. In fact, “most accounts affected had nothing to do with politics at all,” they wrote.
Gadde and Beykpour said that the platform “ranking models take many signals into consideration to best organize tweets for timely relevance.” Twitter’s search engine shows users results from people they find interesting and popular tweets, while ranking lower “tweets from bad-faith actors who intend to manipulate or divide the conversation.”
“Bad-faith actors,” they added, are determined based on how authentic their account appears to be, the actions they take on Twitter and how other users interact with them (for example, how often they are muted, blocked, retweeted or followed).
Gadde and Beykpour said the third criteria may have made it appear that accounts by Republican representatives were being disproportionately affected by the auto-suggestion issue.
“There are communities that try to boost each other’s presence on the platform through coordinated engagement,” they wrote. “We believe these types of actors engaged with the representatives’ accounts– the impact of this coordinated behavior, in combination with our implementation of search auto-suggestions, caused the representatives’ accounts to not show up in auto-suggestions.”
Gadde and Beykpour’s explanation, however, frustrated users on both ends of the political spectrum, who said the company has to do a better job of defining who “bad-faith actors.” Liberals argued (as many have for a long time) that Twitter does not include enough bullies and troll accounts in its definition of “bad-faith actors,” while some conservatives continued to claim that the platform is biased against them.
If what Twitter means by bad faith actors & manipulation is abusers, harassment, hate speech, then say that. Otherwise, this reads as a values-based decision w/opaque criteria for what constitutes “manipulation.” Maybe spell out what specific *behaviors* you’re targeting.
— Robi-Ann Doyle (@doyle1020) July 27, 2018
It says in this article that you judge ‘bad faith actors’ based on who they follow and what they retweet. And then give NO FURTHER INFORMATION on this. That’s *very* dodgy.
— Jamie (@realurbanspice) July 27, 2018
@vijaya a few problems. “How other accounts interact with you(e.g. who mutes you, who follows you, who retweets you, who blocks you, etc)”
If the Left has large preemptive block lists, that creates a problem. (1)
— Lee Doren (@LDoren) July 27, 2018
Yes you do not “shadow ban” in fact you do not ban not even far right trolls accusing people they dislike of being pedophiles. In fact you give them blue check marks and verify them. Just use @Cernovich as an example and that’s that.
— Raúl Vergara (@RaulVB) July 27, 2018