Towards a diachronic understanding of the harm potential of information disorder — reflections on Election 2020

Guest Blogger

From the Workshop on  “News and Information Disorder in the 2020 US Presidential Election,” 

Sam Gill

The study of information disorder is a study in epistemic anxiety. The anxiety that false information may be harmful to our democracy is matched by an anxiety that we have little insight into prevalence and effects of exposure to false information. This, in turn, is matched by an anxiety that digital information platforms are simply funhouse mirrors, distorting and enlarging analog social pathologies. The liturgy then completes its recursive turn in anxiety that we know truly nothing without greater access to platform data, concluding in a supplication to the platform companies. (Or some higher power — for example, a European Union consultative process).

Amen.

Yet what the 2020 election has accentuated is an opportunity for progress, both in assessing the moral harm from information disorder and in measuring it. The key is in identifying the species of harm that are of the highest moral urgency as the starting point for observation. In particular, the material experience of the election suggests that digitally disseminated, accelerated and amplified false information should be understood “diachronically,” to borrow a term from linguistics. That is, the time, place, and manner of false information all matter, as do its provenance and understanding of those most affected. History, in a wide sense, is meaningful for grasping both the mechanics of false information and the magnitude of moral harm.

The first contextual dimension of information disorder underscored by the election is temporal. Timing matters. Seeing, disseminating, and amplifying false information about the election — during a period in which many millions of Americans were making a decision about how to vote — was a clear and present threat to electoral integrity. This would not have been true about precisely the same false information a calendar year earlier. Similarly, while the amplification of false information about the election a year earlier would have been deleterious to the rule of law in a general sense, raising these issues in the aftermath of the election has constituted a graver threat. Today, the rule of law is potentially materially weakened because, it appears, many people have decided — here and now, thanks to digitally accelerated false information — that the recently undertaken democratic procedure was performed illegitimately.

The platform urgency to address COVID-19 misinformation evinces a similar recognition that context matters. Many have noted that, despite some efforts to curb misinformation related to the pandemic, the same patterns of false information continue to obtain traction for other critical health categories, such as vaccines. But, of course, declining vaccination rates had not been seen as a public health crisis — until they became one. COVID-19, however, has been an emergency since its appearance. There is some utility to this distinction. False information about COVID-19, particularly during a period of community spread, is a clear and present danger to public health.

False information in the right place and at the right time vastly amplifies the stakes. This is not to suggest that false information outside of emergent circumstances is benign. To the contrary, social media is not helping us in the quest to eradicate measles; thanks to vaccine-related false information, it’s likely hurting us. It does suggest, however, that the magnitude of harm is larger when timing is relevant.

A second dimension is spatial or contextual, particularly with respect to who is most affected. There is increasing evidence that harmful content of all kinds, including false information about elections, both disproportionately targets — and affects — historically marginalized communities, including women and people of color. This is a more urgent harm for several reasons. First, to adopt Merrill Sanger’s term, false information that targets communities of color, particularly about life- or democracy-critical events like elections, is “syndemic.” This is meant to describe the known biosocial interaction between diseases and the social context of spread. False information about elections targeted toward people of color is syndemic on its face: It amplifies and interacts with the effects of other, known analog deterrents to voting, including various voter suppression techniques. Second, the spatial context of identity introduces new moral harm. That the pattern of false information could be discriminatory is, if accurate, itself objectionable on that basis alone, irrespective of potential disparate impact.

Assuming there is something to a diachronic view, that would suggest a kind of prioritarian argument to address certain kinds of false information first and most aggressively, based their “history” and temporal and spatial context.

This reshapes the question of measurement, with at least two early moves worth considering. The first is to develop a consensus on the types of events more likely to signal magnified harm potential. Pandemics and national elections seem easy. There are likely other low-hanging fruit. An initial but not exhaustive or exclusive framework would be a start.

The second is to improve our systems to understand exposure and influence of critical false information in a way disaggregated by demographic proxies for vulnerable communities. And before we proceed down the important but well-worn path of decrying proprietary control of data, we should also stipulate that companies themselves build this lens into their enforcement efforts and reporting. Airbnb, for example, developed a complex but intriguing system to track potential incidents of race discrimination on its platform through consultation with civil society organizations.

Even if we do proceed down that known and unsatisfying path of noting that the failure to access to platform data is a fatal research flaw, a diachronic, prioritarian lens can illuminate new and potentially productive paths forward. For example, the 16 “critical infrastructure” sectors are subject to unique information policies that enable the uniquely protected sharing of proprietary data so that the Department of Homeland Security can monitor their health. Some have suggested that social media platforms be labeled as “critical infrastructure” to trigger federal cybersecurity protections. Another recent example of such a practice is in the application of fair housing policy, which has been long focused on ameliorating, or preventing, disparate impact in housing. From 2015 to 2018, localities were required to document and track patterns of bias and discrimination in housing in order to meet Affirmatively Furthering Fair Housing obligations.

This diachronic approach to measuring the harm of information disorder, especially with regard to critical events like elections, is merely a sketch meant to raise some useful questions. But, as a conceptual approach, it has a few salutary benefits. First, it can delimit information disorder research by focusing our inquiry on the false information phenomena that matter most. Second, it can accommodate disputes about the relationship between misinformation and other social pathologies by focusing on the syndemic effects of information disorder in a wider sociopolitical context. Third, it can clarify and narrow the claims to privately-controlled data, locating them in more specific arguments about critical or crisis moments.

This approach is unlikely to satisfy our descriptive, social scientific pieties. But it may help us to address the normative exigencies that are moving ahead whether we have scientific clarity or not. Consider the responses of the major social media platforms, particularly Facebook, Twitter, and YouTube, in both the lead-up to Election Day 2020 and during the weeks following. They can rightly be characterized as inconsistent, halting, and inconstant. They can also be described as frenetic and urgent.

The race to thwart the problem is on — with or without fundamental understanding.

Sam Gill is senior vice president and chief program officer at the John S. and James L. Knight Foundation.

Cross posted at the Knight Foundation