Adrian K. Yee

Machine learning models of misinformation detection are increasingly being used and yet their risks have been under analyzed, requiring insights from philosophy of science and political philosophy. A taxonomy of types of risks and simple models estimating them is provided that is sensitive to the value-laden features of judgments of misinformation and weighted by a potential item of misinformation’s impact on respective stakeholders. Failing to do so is incompatible with a variety of civil virtues that are desired by not only liberal democratic societies but authoritarian ones as well, suggesting the general applicability of this taxonomy of risks to contemporary and future societies.

Extant accounts of multidimensional poverty have continued to neglect the extent at which severe deprivations in adequate information usage and interpretation are a legitimate feature of poverty. I propose a new hermeneutic theory of information poverty that combines the insights from philosophical and development economic theories, is complementary to extant multidimensional poverty metrics, and is illustrated with a detailed case study concerning “persons with albinism” in contemporary Tanzania. I argue that this theory allows for partially measurable features of information poverty, is applicable in policy contexts, and should be of concern to, among others, advocates of the capabilities approach in political philosophy.

Empirical social sciences routinely model misinformation as exhibiting dynamics analogous to vaccinable diseases or contagious outbreaks, as in inoculation theory and other epidemiological models. However, idiosyncratic features of the social construction of misinformation violate the biological analogy in significant ways, rendering these models far weaker in effect size, predictive accuracy, and explanatory power than has been claimed. Four arguments are discussed regarding problems with the ontology of misinformation posited in these models, methods for measuring misinformation, individuation of mechanisms, and application of interventions. A conclusion is drawn that model transfer from biology has often been unwarranted in misinformation studies and that alternative methods should be pursued instead.

Judgments of misinformation are made relative to the informational preferences of the communities making them. However, informational standards change over time, inducing distribution shifts that threaten the adequacy of machine learning models of misinformation. After articulating five kinds of distribution shifts, three solutions for enhancing success are discussed: larger static training sets, social engineering, and dynamic sampling. I argue that given the idiosyncratic ontology of misinformation, the first option is inadequate, the second is unethical, and thus the third is superior. However, I conclude that the prospects for machine learning models of misinformation are far weaker than most have presupposed, given that both epistemic and non-epistemic values are difficult to operationalize dynamically in machine code, rendering them surprisingly at most a species of recommender systems rather than literal truth detectors.

Governments and social scientists are increasingly developing machine learning methods to automate the process of identifying terrorists in real time and predict future attacks. However, current operationalizations of “terrorist”’ in artificial intelligence are difficult to justify given three issues that remain neglected: insufficient construct legitimacy, insufficient criterion validity, and insufficient construct validity. I conclude that machine learning methods should be at most used for the identification of singular individuals deemed terrorists and not for identifying possible terrorists from some more general class, nor to predict terrorist attacks more broadly, given intolerably high risks that result from such approaches.

Francis Ysidro Edgeworth’s unduly neglected monograph New and Old Methods of Ethics (1877) advances a highly sophisticated and mathematized account of social well-being in the utilitarian tradition of his 19th-century contemporaries. This article illustrates how his usage of the ‘calculus of variations’ was combined with findings from empirical psychology and economic theory to construct a consequentialist axiological framework. A conclusion is drawn that Edgeworth is a methodological predecessor to several important methods, ideas, and issues that continue to be discussed in contemporary social well-being studies.

Current methods of operationalizing concepts of misinformation in machine learning are often problematic given idiosyncrasies in their success conditions compared to other models employed in the natural and social sciences. The intrinsic value-ladenness of misinformation and the dynamic relationship between citizens' and social scientists' concepts of misinformation jointly suggest that both the construct legitimacy and the construct validity of these models needs to be assessed via more democratic criteria than has previously been recognized.

There have recently been novel applications of medical systematic review guidelines to economic policy interventions which contain controversial methodological assumptions that require further scrutiny. A landmark 2017 Cochrane review of unconditional cash transfer (UCT) studies, based on the Grading of Recommendations Assessment, Development and Evaluation (GRADE), exemplifies both the possibilities and limitations of applying medical systematic review guidelines to UCT and universal basic income (UBI) studies. Recognizing the need to upgrade GRADE to incorporate the differences between medical and policy interventions, the GRADE Public Health Project Group (PHPG) was convened to enumerate and address these methodological challenges. However, in light of our analysis of additional methodological challenges that arise for UCT and UBI studies, we argue that the adaptation of medical systematic review guidelines to economic methodology is far from straightforward and is in fact more challenging than claimed by the PHPG.

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.

The history of economic thought witnessed several prominent economists who took seriously models and concepts in physics for the elucidation and prediction of economic phenomena. Econophysics is an emerging discipline at the intersection of heterodox economics and the physics of complex systems, with practitioners typically engaged in two overlapping but distinct methodological programs. The first is to export mathematical methods used in physics for the purposes of studying economic phenomena. The second is to export mechanisms in physics into economics. A conclusion is drawn that physics transfer is often justified at the level of mathematical transfer but unjustified at the level of mechanistic transfer.