A study carried out by researchers at Arizona State University, Texas A&M and Yahoo, funded in part by the US Military's Office of Naval Research, looked at 2,686 Twitter posts to create a system that reliably spots future online protesters.
Twitter better go ahead and launch that algorithm thing now so they can algorithm all of this dissent out of our timelines.
— Chris M (@aryst0krat) February 6, 2016
This was no small feat, as online behavior opens only a small window into a person's offline thinking — and possibly dissident feelings.
"The ways in which protest-related events affect a person are not observable, resulting in a lack of knowledge of factors operating at that time causing his next post to be a declaration of protest," the researchers and authors of the study explained.
"A User is subject to various types of influence in his past, and many of them are in conflict with each other. This may lead to ambiguities on whether his posts will contain declarations of protest in the future," they added.
The researchers sifted through 2,686 Twitter posts regarding last year's Nigerian general election, which eventually sparked widespread protests due to irregularities and militant violence. What they realized is that the main way of predicting whether a Twitter user will partake in future protest movements is his or her online interaction with people who are already involved in protesting.
Don't go there, twitter. We need a platform that isn't weakened by an #algorithm. Not perfect but it is enough to voice dissent! #RIPTwitter
— Eluned Francis (@lunaed) February 6, 2016
Specifically, the likelihood of a user joining a protest increases the more he or she is mentioned in tweets related to the protest by other users who are already protesting.
An algorithm designed following these principles was tested on the Nigerian tweets, and it accurately predicted if users would tweet protest posts. Interestingly, the model was based on the Brownian motion theory, which studies how particles of fluid move around and interact with each other.
What are the applications of this model? According to the paper, "predicting if the next post of a given user will be a declaration of protest will help in estimating the number of protest participants, facilitating better protest organization".
Still, it is not far-fetched to think the system could be used to keep a tab on protest movements, rather than facilitating them. Military news website Defense One, for instance, mused about how it could be deployed against Daesh, also known as ISIL — although the scientists underlined that protest and radicalization are two different phenomena.
What is more worrying is that in the wrong hands, the algorithm would give governments an upper leg over political dissidents they do not like — a sort of pre-cog capability that would leverage Twitter to single out and target protesters before they even start protesting.