The Social Algorithm Behind Democracy’s Decay
Photo by Antonio Batinić
By CAMERON GIL
For years, the public sphere has been littered with anecdotes and news stories concerning the sometimes drastic changes in behavior of typical members of society - our family members, neighbors, coworkers, fellow voters, elected officials who design the systems of government, and bureaucrats tasked with upholding our government.
The concession of a republic is that all members within implicitly partake in the republic’s evolution and expression. In short: “We the People.” However, tides of polarization, exclusionary acts, and fierce political competition have shifted views of the republic’s composition away from ‘those within’ to ‘ours are the republic.’ This is political philosophy, yet, the empirical evidence across social media studies gathered by our research team at the Organization for Social Media Safety indicates there are social media platform underpinnings and algorithms which exacerbate a distancing of those within the republic.
One of the most worrying signals for democracy and freedom is research that finds viewing the “other side” emboldens division. More intensive data collection over the years, from probability surveys (Lee et al. 2014) to in-depth surveys and tracking engagement habits of those exposed to opposing views (Gillani et al. 2018), seems to support the view that social media platforms are changing to embolden division or that the behavioral mechanics are already seeded among digital stakeholders. One of the more stark pieces of evidence that social media platforms are emboldening a bubble, promoting in-group seeking and reinforcing behaviors (Boutyline and Willer 2016), in combination with making people inside a bubble think they are free from ideological possession indicates that interjecting diverse views in one’s feed can make one more likely to decrease in their belief of being in a political bubble and decrease willingness to engage with those outside the bubble (Gillani et al. 2018). The worry for those monitoring these developments returns to how digital behavior accelerates and what it means for the “real world.” Does anonymity and digital distance promote caricatures of other participants in democracy?
Negative feedback loops persist inside the bubbles of social media platforms, particularly those with algorithms which prioritize engagement (likes, comments, messages, etc.), in large part because of the algorithm rewarding the intensity of negative emotions and divisive behavior. Content ripe for abusing these platforms’ code consists of politically divisive, exclusionary content that goads negative sentiments which materialize as longer posts, higher quantity of comments, and group feedback too (Bessi 2015; Zollo et al. 2015). The validity of antidemocratic information and conspiracy theories is one that is easily overlooked when individuals and groups feel attacked or sidelined. Rather than engaging with the legitimacy of the source information, folks enter “flame wars” and “ratio’ing” to dogpile on the other side. As we’ve seen across the data and research, this engagement outside the bubble often pushes people apart further without negative emotions - the heat of the argument exacerbates this behavior further.
There is a line of research into solutions for these social dynamics propagated by social media. For example, Information Credibility on Twitter (Gupta 2014) proposes a machine-learning approach to sort out what, and who, is a valid source of information in real time. However, the underlying methods are themselves open to scrutiny for form and bias. Furthermore, the judgment of a machine is particularly at threat by those seeking to divide and discredit others from legitimate democratic discourse. To those within the in-group seeking to extirpate others from the republic, to inside the bubble, is not a machine’s voice a modern strawman? Or is the value of technological interventions in the digital world more as a buffer against the expansion of division. To preserve our collective view of what freedom means, it may recommend more than a quick peek under the hood of validity, it may necessitate a wholesale inspection into how the source code of our social platforms encode into us undemocratic behaviors.
As many who peruse online forums and social media know, this behavior has seeped into the halls of government. Elected representatives are not immune from replicating and accelerating the behaviors noted in the research. The social media presence and audience characteristics of politicians reflects calcification of echo chamber bubbles seen across social media platforms (Hong and Kim 2018). Take the behaviors and amplify them with the political arena where repetitive success at the ballot box is largely the primary motivating factor for a representative. If gerrymandering of districts creates polarization, then as we’ve seen in the empirical data, gerrymandering of social media bubbles further complicates the picture through the in and out group divisions targeted at who is a legitimate democratic participant. If those writing and interpreting the laws which encode freedom and democracy into the republic are participants in the mechanism of social media-based division, does freedom end where the bubble begins?
The systemic feedback of negativity reinforces the bubble, and all the bubble distorts one’s vision so those inside see the outside as unfree or even as a threat - that extends to democratic practices and norms. As we see in this slice of studies, the dynamics that persist in social media platforms help the more divisive elements of social dynamics thrive in the digital world. The systemic feedback of negativity reinforces the bubble that cages people in their slice of the digital world, and worse, threatens to divide democracy at the most fundamental level - “We the People.”
There is much more to be done as the digital world, in all its value and infamy, becomes more tangible. The Organization for Social Media will continue to track and catalog an extensive database of relevant research, data-informed recommendations to governments, non-government actors, and at forums including the upcoming IX Summit of the Americas Thematic Subregional Working Groups for Health, Digital Transformation, and Democratic Governance.
___________________
Cameron Gil is a member of the Pacific Council and is Director of Student Advisory Councils at the Organization for Social Media Safety.
The views and opinions expressed here are those of the author and do not necessarily reflect the official policy or position of the Pacific Council.
