Everyone knows that Twitter is a great tool to spread any idea as long as you have the necessary means. Now a recent study prepared by the platform itself has concluded that your amplification algorithm has more in common with the political right than meets the eye.
When we enter Twitter, we have two options to view our timeline. On the one hand, through an algorithm that recommends certain content based on what we follow and have interacted with, and on the other hand, there is seeing the chronology through the most recent tweets. As they affirm from the platform, The reason for this study was to deepen the better understanding of this algorithm.
The study has been carried out by its ethics, transparency and responsibility team through machine learning, known as META. The team led by Rumman Chowdhury analyzed millions of tweets from elected officials in up to seven countries., which include Spain. The investigation also collected data on millions of tweets that contained links to other media.
As Chowdhury explains on the official Twitter blog, in all countries except Germany, those tweets from the political right were mostly amplified by the platform’s algorithm. In addition, the same thing happened with those tweets that contained links to media on the right, always being tweets outside the media in question.
Twitter also concluded that the effects of the algorithm did not have to affect individuals from the same political party in the same way, since they claim, ideology is not a factor included in their systems.
The algorithm amplifies the user’s own actions
As they comment from the platform, to begin this investigation three questions were asked. The first of them referred to the amount of amplification in political content that there is when we order the tweets through the algorithm compared to when we order them by most recent. They also wondered if the algorithm amplified any particular type of political group more than another.. The same was asked in relation to the news media themselves.
The classification of the media themselves according to their ideology was not carried out by Twitter, but rather went to a third party investigation. Delving into the reason for the matter, Chowdhury concluded in an interview for the Protocol medium that part of this amplification “could have been driven by the user himself”, which means that it would be related to how people use this platform.
From the official blog itself they have assured that they will examine this case to “mitigate any lack of equity” that this may cause. Nevertheless, have pointed out that their algorithm is not the cause of the problem, since the algorithm’s own function is to amplify without having been altered or having included any preference at the time of its construction.
Via | Twitter