November 11, 2016 Last Updated 9:16 am

Study reports that nearly 20% of election-related tweets were ‘algorithmically driven’

The influence of both foreign generated social media content, and that of ‘bots’, will likely be taken more seriously by publishers who have too often have been in denial

A study, released the day before the election (and written about by MIT Technology Review on Election Day) comes to the conclusion that a major share of tweets regarding the upcoming election were coming from bots – “algorithmically driven entities that on the surface appear as legitimate users.”

“By leveraging state-of-the-art social bot detection algorithms, we uncovered a large fraction of user population that may not be human, accounting for a significant portion of generated content (about one-fifth of the entire conversation),” the authors, authors Alessandro Bessi and Emilio Ferrara of the University of Southern California Information Sciences Institute, wrote. “We inferred political partisanships from hashtag adoption, for both humans and bots, and studied spatio-temporal communication, political support dynamics, and influence mechanisms by discovering the level of network embeddedness of the bots. Our findings suggest that the presence of social media bots can indeed negatively affect democratic political discussion rather than improving it, which in turn can potentially alter public opinion and endanger the integrity of the Presidential election.”

study-table4One of the many charts and grapsh including in the study by Alessandro Bessi and Emilio Ferrara

I highly recommend Nanette Byrnes’s story here, as well as the study, to be found here. This has been an issue that has been front and center here at TNM for the entire year. Two major developments have had an enormous impact on our world in recent years – the growth of social media (and content consumption on social media sites), and the introduction of automated, or foreign, generated content into that ecosystem.

For the past couple years TNM has monitored the influx of foreign based commenters who masquerade as domestic readers. Here, this comes in the form of comment spam, always deleted and banned based on their IP address. But TNM has reached out to many media properties to understand what they are seeing, and how they are handling it. The answers I have gotten back are most depressing: they are not, or they are in denial.

Byrnes interviewed the author of the story, who made an interesting claim: “We find in some states, in particular in the South and the Midwest, there are way more bots than in any other state. People might think there is a real grassroots support there, but in reality it’s all generated by the bots.”

Of course, when the votes were counted, both regions went for Trump. One could interpret this in one of two ways: they mistakenly identified tweets as bots, when they were really Trump supporters; or that these regions were being targeted for political reasons.

It should be noted, and the authors do themselves, that accurately identifying bot traffic can be difficult. This may explain why the author’s graphs showing bot traffic and human traffic are nearly identical. But one state seems to stand out in generating bot tweets: Georgia, which generated the most bot traffic according to the authors of the study.

This is an important study, and thank you to MIT Technology Review for finding it.

Comments are closed.