Twitter’s algorithms prefer to shove right-wing content down users’ throats
The platform is working to understand why that is happening.
Twitter recently took to a blog post on the platform’s website to share findings from internal research that the company has been performing. The company’s research found that its algorithms that are used to display relevant content tend to prefer showing political content from right-wing sources as opposed to content from the political left.
In the blog post, Twitter detailed the steps that it took to perform this research. The company analyzed millions of tweets from accounts in seven different countries, studying how its algorithms amplified certain content. It also looked at hundreds of millions of tweets that included links to news outlets.
The research showed that in six out of the seven countries researched, all but one saw tweets posted from the political right amplified on users’ timelines more than similar posts from the left. It also found that right-leaning news outlets tend to see more amplification when compared to more left-leaning outlets.
Twitter doesn’t seem to know exactly why this is happening this way. Rumman Chowdhury, head of machine learning at Twitter, told Protocol, “When algorithms get put out into the world…We can’t model for how individuals or groups of people will use Twitter, what will happen in the world in a way that will impact how people use Twitter.”
While it’s not very reassuring to learn that the platform doesn’t know why its algorithms act the way they do, it’s nice to see this kind of transparency from Twitter. It’s a breath of fresh air after hordes of Facebook internal research had to be leaked before the company would share anything.
Twitter also said that it is working on a way to make the data it used for this research available to third-party researchers so that these findings can be validated.
Of course, user privacy is the main concern with something like this, but the platform says it is “finalizing a partnership to leverage privacy-preserving technology to enable third-party researchers to reproduce this type of work, while also protecting and safeguarding the privacy of people who use Twitter.”
- Facebook is changing its name – here’s what Twitter thinks it should be
- YouTube’s updated misinformation policies led to less bullshit on Facebook and Twitter
- Twitter is testing out ads in replies because ads every five tweets wasn’t enough
- Instagram’s solution to its teen problem? A new “take a break” feature