Trump Twitter attack: What exactly is online shadow banning?

27 Jul 2018

Twitter logo. Image: Tero Vesalainen/Shutterstock

US president Donald Trump accused Twitter of shadow-banning Republicans, but what does that mean?

Earlier in the week, US president Donald Trump accused Twitter of shadow-banning the accounts of prominent Republican politicians. An article from Vice News sparked the reaction from Trump, with the piece alleging that Twitter was limiting the visibility of Republicans in search results, framed as shadow banning without sufficient evidence.

Trump then accused the platform of illegal discrimination, and outlets such as Breitbart and InfoWars reported the shadow-banning story as fact.

What is shadow banning?

In short, shadow banning is deliberately making content from an account undiscoverable to everyone except the person who posted it, without the poster knowing.

The idea is to make harmful content invisible and eventually persuade the offending user to willingly leave the platform when their engagement falls off a cliff. Reddit used shadow banning as its only form of banning for years and it was used on the platform until the end of 2015. Some proponents of the practice say outright bans will see the user set up another account and continue bad behaviour.

Twitter responds

Twitter’s Vijaya Gadde and Kayvon Beykpour said: “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.”

According to the company, the reduced visibility of accounts in auto-suggested searches was simply a bug affecting many accounts, not just those affiliated with the Republican party. “As mentioned above, the issue was broad-ranging and not limited to political accounts or specific geographies. And most accounts affected had nothing to do with politics at all.”

More work to be done

The company said it had remedied the problem and would continue to examine its system for similar errors.

According to Twitter, the platform’s ranking models take different signals into account to best organise tweets in terms of relevance and timeliness. It ranks tweets from ‘bad faith’ actors who aim to manipulate the discourse. These users are determined based on how authentic their account seems to be as well as how often other users follow, mute or retweet the account.

Critics on both sides of the political spectrum said Twitter needs to better define what makes a bad-faith actor. Both conservative and more liberal users have their issues, with some saying many troll accounts remain live on the platform, while others feel they are being silenced – something the company has always refuted.

Twitter logo. Image: Tero Vesalainen/Shutterstock

Ellen Tannam was a journalist with Silicon Republic, covering all manner of business and tech subjects