Within the spectrum of social networks, when you think about the consumption of news or other current material, the preferred alternative is usually Twitter.
Within this platform, the topics that set trends within the debate at a local or global level are grouped in an outstanding ranking that, according to recent research, is highly vulnerable to external manipulations.
Twitter trending algorithm under the microscope
An investigation carried out at the Swiss Federal Institute of Technology in Lausanne (EPFL), pointed out that the social network does not verify if a tweet has been deleted when evaluating which keywords will enter the trends section. By bypassing this filter, one or more coordinated accounts can inflate these statistics, adding mentions of certain keywords with deleted tweets, which can be repeated cyclically.
This new intervention technique, defined as “ephemeral autosurfing”, has been put into practice on Twitter since 2015, according to the researchers, who accuse that its influence has reached local and global scales.
The report that compiles the reports of this study, indicates that at least 20% of the global trends generated during the period they observed, had their origin through this type of attack.
The investigation pointed out that this vulnerability already enjoys a certain popularity, and there is even a whole business model around this class of shares.
The influence of Twitter and what this vulnerability implies
Platforms such as Twitter, as well as the media and any other instance of a similar nature, have a strong influence on the social imaginary. For this reason, what this investigation accuses is not a minor matter.
The Next Web echoed this report, noting that they contacted Twitter to grant their right to reply, which so far has not been answered.
The EPFL report, in addition to discussing the security implications that this situation generates on the platform, along with its impact on its users and society, also proposes an action plan to strengthen the defense of Twitter.
This document, published as paper, also throws its darts at the studied social network, claiming that they have ignored the warnings shared.
“Due to the severity of the attack, we notified Twitter (once in July 2019 and again in June 2020) and always a detailed description of the attack and the accounts involved. After the first notification, they acknowledged that the attacks exist (July 2019) and after the second notification (June 2020), they replied that they would forward the message to the appropriate team to address it. We have continued since then, but have not received any signs that they are progressing. Attacks on Twitter trends continued in February 2021 “, points out the document in its introductory section.