Twitter has been used many times to manipulate information.
Such manipulation comes both from humans and bots.
Some cases got significatn media attention like the United States presidential election of 2016, but others might have slipped unnoticed.
Some examples of Russian government-linked organization known as the Internet Research Agency (IRA).
We’ve seen fascinating examples like from @fs0c131y with very well-made graphical explanations of the manipulation.
You can check all the 28 tweets in one here.
Pulled 5.5K recent tweets which include #WeSupportPolandSummit@geoffgolberg now deleted Tweet
6.3K accounts tweeted, retweeted, or were mentioned in said tweets
The hashtag is clearly being gamed cc: @benimmopic.twitter.com/iwPhpTmEDe— Geoff Golberg (@geoffgolberg) January 17, 2019
Ben Nimmo, a former Journalist, now working for the Defense at the Atlantic Council’s Digital Forensic Research Lab (DFRLab) is known to be outstanding in analyzing this kind of manipulation.
He’s been attacked by the Kremlin propaganda outlets threatened by the far right and declared dead by bot herders.
Ben Nimmo proposes trough the Oxford Project on Computational Propaganda Workpaper to use a computational method to calculate the extent to which a given flow of Twitter traffic has been subject to manipulation called CMT.
This formula allows comparing different traffic flows against measurable criteria and see which flows appear to have been most subject to manipulation.
The paper goes and compares different phrases and hashtags from Twitter to clearly identify how advanced was a manipulation.
From the paper:
“The lower group of seven control samples clusters together, with a variation of just over four points between the lowest and highest. The upper group, consisting of the two manipulated Gulf samples, begins 1.9 points higher. This suggests that, even in larger and less obviously manipulated cases, the CTM can give a reliable indication of the degree of manipulation to which a traffic flow has been subjected.”
For security researchers, this could be a new and standardized way that could be used to compare and define the magnitude of Twitter disinformation campaigns.