As I discussed on Thursday, Twitter found itself in hot water after the left-wing news outlet Vice News reported that Republicans were “shadow banned” while Democrats were not prompting a response.
Twitter CEO Jack Dorsey took his social media platform to address the situation.
“A short thread addressing some issues folks are encountering as a result of our conversational health work, specifically the perception of ‘shadowbanning’ based on content or ideology. It suffices to say we have a lot more work to do to earn people’s trust on how we work,” he tweeted.
A short thread addressing some issues folks are encountering as a result of our conversational health work, specifically the perception of “shadowbanning” based on content or ideology. It suffices to say we have a lot more work to do to earn people’s trust on how we work. https://t.co/MN97l7w7RF
— jack (@jack) July 25, 2018
Dorsey quoted a tweet from Kayvon Beykpour, the Twitter project lead and co-founder of Periscope, who said. “To be clear, our behavioral ranking doesn’t make judgements [sic] based on political views or the substance of tweets.”
What is their behavioral ranking? We don’t know. Beykpour along with Vijaya Gadde, the Legal, Policy and Trust & Safety Lead for Twitter co-wrote an article on the social media platform’s blog:
We do rank tweets and search results. We do this because Twitter is most useful when it’s immediately relevant. These ranking models take many signals into consideration to best organize tweets for timely relevance. We must also address bad-faith actors who intend to manipulate or detract from healthy conversation.
As a specific example, if a search result has 30,000 tweets, here’s what we take into consideration when ranking:
- Tweets from people you’re interested in should be ranked highly
- Tweets that are popular are likely to be interesting and should be higher ranked
- Tweets from bad-faith actors who intend to manipulate or divide the conversation should be ranked lower
They explained the last bullet point:
This last bullet is the basis of our work around serving healthy public conversation. Here are some of the signals we use to determine bad-faith actors:
- Specific account properties that indicate authenticity (e.g. whether you have a confirmed email address, how recently your account was created, whether you uploaded a profile image, etc)
- What actions you take on Twitter (e.g. who you follow, who you retweet, etc)
- How other accounts interact with you (e.g. who mutes you, who follows you, who retweets you, who blocks you, etc)
They do this Beykpour explained through “machine learning.”
“In May, we started using behavioral signals and machine learning to reduce people’s ability to detract from healthy public conversation on Twitter. This approach looks at account behavior & interactions with other accounts that violate our rules,” he tweeted.
On 2) Some accounts weren’t being auto-suggested even when people were searching for their specific name. Our usage of the behavior signals within search was causing this to happen & making search results seem inaccurate. We’re making a change today that will improve this.
— Kayvon Beykpour (@kayvz) July 25, 2018
Twitter said the issue was not limited to Republicans, but included Democrats.
“Yes, some Democratic politicians were not properly showing up within search auto-suggestions as result of this issue. 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,” Beykpour and Gadde wrote.
Why were the Republican members of Congress impacted, “For the most part, we believe the issue had more to do with how other people were interacting with these representatives’ accounts than the accounts themselves,” they added.
Twitter’s response is just obfuscation and leaves us with more questions than answers.
- Why does Twitter imply that machine learning is somehow neutral? It’s not. It does what the programmers program it to do.
- How does Twitter determine someone is trying to “manipulate” and “divide” a conversation.
- How do they determine “conversational health” what does that even mean? Do they not recognize that phrase sounds big brother-ish and creepy, not to mention, vague to users?
- Why should who you follow and those who follow you determine whether or not you are acting in bad faith?
- Do they not realize how other accounts interact with one’s account is subject and out of the control of a Twitter user?
- Why did the behavior issue flag specific accounts and not others?
- What did Twitter do to fix this?
- They said the problem also impacted Democrats, ok, who? Name names. Vice News did not find that to be true and it’s not like they are trying to gin up a fake controversy for Republicans.
This “shadow banning” controversy is just one issue of several. Take, for instance, tweets that are initially blocked from view because they may contain “potentially sensitive material.” Are we talking pornography? Gore? Violence? Vulgar language?
Perhaps it does address that, but it also addresses content liberals may not like. Case in point from a tweet I saw Friday morning:
What was this potentially sensitive content?
You’ve got to be kidding me.
Then there has been a noticeable disparity between the suspension and de-verifying of Twitter users on the right (and, no, not just alt-right users) and those on the left. Federal Communications Commission Chairman Ajit Pai noted this double standard.
“Now look, I love Twitter and anyone who knows me knows that I use it all the time. But let’s not kid ourselves, when it comes to a free and open Internet, Twitter is part of the problem. The company has a viewpoint and uses that viewpoint to discriminate,” Pai said in a speech at R Street Institute last fall.
“To say the least, the company appears to have a double standard when it comes to suspending or de-verifying conservative users’ accounts as opposed to those of liberal users. This conduct is many things, but it isn’t fighting for an open internet,” he added.
So, yeah, Dorsey is right that Twitter has “a lot more work to do to earn people’s trust” and they can start with a straightforward and truthful explanation about what happened this week.
Until they do, they’ll likely continue to see news like what the Associated Press tweeted out Friday afternoon.
BREAKING: Twitter plunges 20.5 percent after reporting decline in monthly users, second-biggest loss since going public in 2013.
— The Associated Press (@AP) July 27, 2018