EPFL Discover Twitter Trends Manipulation on Large Scale

Jun, 2021 - By WMR

Nearly 50% of local trending topics in Turkey on Twitter are fake, according to the EPFL scientists.

New research by the Swiss EPFL (Federal Institute of Technology Lausanne) has discovered that nearly 50% of local trending topics in Turkey on Twitter are fake, a manipulation scale previously unheard. It even for the very first time proves that various trends are solely created by bots, owing to a loophole in algorithm of Twitter's Trends. Social media has changed way of connecting and interacting of people in unimaginable ways.

One influencing social media mechanism, for instance, is "Twitter Trends." This platform makes use of algorithm for determining topics that are hashtag-driven, that gain popularity at a certain point of time, alerting users on twitter to top phrases, subjects, words, along with popular hashtags locally and globally. Now emphasizing on Turkey, new EPFL study, from the Distributed Information Systems Laboratory, part of School of Computer and Communication Sciences, noticed a vulnerability within the algorithm which decides Trending Topics on Twitter - it doesn’t count the deletions. This loophole enables attackers to drive trends of their choice to Twitter Trends’ top despite shortly afterwards deleting the tweets containing this candidate trend . The discovered fake trends consists of phishing apps, political slogans, disinformation campaigns, gambling promotions, hate speech against vulnerable populace and even wedding proposals.

Rebekah Overdorf, one of the authors of this study said that such manipulation has severe implications since Twitter Trends easily gets the world’s attention. These trends are reported by Broader media channels, that use it as proxy for what the public is talking about, however this is unfortunately a manipulated proxy, that distorts the view of people about the kind of conversations actually going on. The researchers reached twice to Twitter, with the micro blogging site acknowledging the vulnerability in Trends algorithm in both scenarios its. Twitter refused for making any changes in first case, the company in second case didn’t respond to follow-up emails from researchers.