Information Deprivation and Democratic Engagement

Today's post is by Adrian K. Yee. Adrian is finishing his PhD at the University of Toronto focusing on the intersection of philosophy of science, politics, and economics (PPE) and will begin a position as Research Assistant Professor at Lingnan University starting August 2023. 

Adrian has previously published on ‘econophysics’ (applications of physics to economic & financial phenomena) and misinformation studies, and has a forthcoming paper improving the methodology of Universal Basic Income studies. His upcoming research projects focus on philosophy of AI, attention economics, and epistemological & ethical issues in military intelligence analysis.

Adrian K. Yee

In the paper ‘Information Deprivation and Democratic Engagement’, I argue that there remains no consensus among social scientists as to how to measure and understand forms of information deprivation such as misinformation. Machine learning and statistical analyses of information deprivation typically contain problematic operationalizations which are too often biased towards epistemic elites’ conceptions that can undermine their empirical adequacy. A mature science of information deprivation should include considerable citizen involvement that is sensitive to the value-ladenness of information quality and that doing so may improve the predictive and explanatory power of extant models.

There are three central problems extant models face. Firstly, operationalizations of misinformation are too focused on problematic alethic accounts (i.e. misinformation defined as simply false information) considering there is false information that is extremely useful and explanatorily powerful (e.g. any regression method in applied statistics) and there is true information that is misinformation (e.g. spin, malinformation, misleadingness etc.). 

Secondly, the intrinsic value-ladenness of judgments of misinformation entails that determining information quality is not something intersubjectively verifiable in the way natural scientific phenomena are; hence, information quality is intrinsically socially constructed and negotiable by relevant stakeholders. 

Thirdly, it follows that if we wish to avoid epistocracies, or harmful misinformation laws like Singapore’s since 2019, which fail to provide an adequate definition of misinformation (i.e. it simply allows the state to fine or imprison those it considers engaged in seditious informational practices), then we require further democratic engagement in the adjudication of information quality. As it stands, epistemic elites (e.g. journalists, university researchers, government policymakers, etc.) are almost entirely the ones who are solicited for adjudicating information quality, especially in supervised learning contexts in algorithmic content curation. 

And yet, given that information quality is manifestly not merely a matter of truth, but can involve considerations of explanatory power, relevance, aesthetic quality, moral consequences, parsimony, and other epistemic and non-epistemic values, it follows that such values will require input from average citizens much more so than is currently the case in extant models.