Journal of Political Philosophy, Volume 30, Issue 2, Page 209-229, June 2022.

It is becoming more common that the decision-makers in private and public institutions are predictive algorithmic systems, not humans. This article argues that relying on algorithmic systems is procedurally unjust in contexts involving background conditions of structural injustice. Under such nonideal conditions, algorithmic systems, if left to their own devices, cannot meet a necessary condition of procedural justice, because they fail to provide a sufficiently nuanced model of which cases count as relevantly similar. Resolving this problem requires deliberative capacities uniquely (...) available to human agents. After exploring the limitations of existing formal algorithmic fairness strategies, the article argues that procedural justice requires that human agents relying wholly or in part on algorithmic systems proceed with caution: by avoiding doxastic negligence about algorithmic outputs, by exercising deliberative capacities when making similarity judgments, and by suspending belief and gathering additional information in light of higher-order uncertainty. (shrink)

This article considers the issue of opacity as a problem for socially consequential mechanisms of classification and ranking, such as spam filters, credit card fraud detection, search engines, news trends, market segmentation and advertising, insurance or loan qualification, and credit scoring. These mechanisms of classification all frequently rely on computational algorithms, and in many cases on machine learning algorithms to do this work. In this article, I draw a distinction between three forms of opacity: opacity as intentional corporate or state (...) secrecy, opacity as technical illiteracy, and an opacity that arises from the characteristics of machine learning algorithms and the scale required to apply them usefully. The analysis in this article gets inside the algorithms themselves. I cite existing literatures in computer science, known industry practices, and do some testing and manipulation of code as a form of lightweight code audit. I argue that recognizing the distinct forms of opacity that may be coming into play in a given application is a key to determining which of a variety of technical and non-technical solutions could help to prevent harm. (shrink)

I use a theorem from machine learning, called the “No Free Lunch” theorem to support the claim that non-epistemic values are essential to theory choice. I argue that NFL entails that predictive accuracy is insufficient to favor a given theory over others, and that NFL challenges our ability to give a purely epistemic justification for using other traditional epistemic virtues in theory choice. In addition, I argue that the natural way to overcome NFL’s challenge is to use non-epistemic values. If (...) my argument holds, non-epistemic values are entangled in theory choice regardless of human limitations and regardless of the subject matter. Thereby, my argument overcomes objections to the main lines of argument revealing the role of values in theory choice. At the end of the paper, I argue that, contrary to common conception, the epistemic challenge arising from NFL is distinct from Hume’s problem of induction and other forms of underdetermination. (shrink)

Often machine learning programs inherit social patterns reflected in their training data without any directed effort by programmers to include such biases. Computer scientists call this algorithmic bias. This paper explores the relationship between machine bias and human cognitive bias. In it, I argue similarities between algorithmic and cognitive biases indicate a disconcerting sense in which sources of bias emerge out of seemingly innocuous patterns of information processing. The emergent nature of this bias obscures the existence of the bias itself, (...) making it difficult to identify, mitigate, or evaluate using standard resources in epistemology and ethics. I demonstrate these points in the case of mitigation techniques by presenting what I call ‘the Proxy Problem’. One reason biases resist revision is that they rely on proxy attributes, seemingly innocuous attributes that correlate with socially-sensitive attributes, serving as proxies for the socially-sensitive attributes themselves. I argue that in both human and algorithmic domains, this problem presents a common dilemma for mitigation: attempts to discourage reliance on proxy attributes risk a tradeoff with judgement accuracy. This problem, I contend, admits of no purely algorithmic solution. (shrink)

In analyzing oppressive systems like racism, social theorists have articulated accounts of the dynamic interaction and mutual dependence between psychological components, such as individuals’ patterns of thought and action, and social components, such as formal institutions and informal interactions. We argue for the further inclusion of physical components, such as material artifacts and spatial environments. Drawing on socially situated and ecologically embedded approaches in the cognitive sciences, we argue that physical components of racism are not only shaped by, but also (...) shape psychological and social components of racism. Indeed, while our initial focus is on racism and racist things, we contend that our framework is also applicable to other oppressive systems, including sexism, classism, and ableism. This is because racist things are part of a broader class of oppressive things, which are material artifacts and spatial environments that are in congruence with an oppressive system. (shrink)

Scientists depend on complex computational systems that are often ineliminably opaque, to the detriment of our ability to give scientific explanations and detect artifacts. Some philosophers have s...

According to the Value-Neutrality Thesis, technology is morally and politically neutral, neither good nor bad. A knife may be put to bad use to murder an innocent person or to good use to peel an apple for a starving person, but the knife itself is a mere instrument, not a proper subject for moral or political evaluation. While contemporary philosophers of technology widely reject the VNT, it remains unclear whether claims about values in technology are just a figure of speech (...) or nontrivial empirical claims with genuine factual content and real-world implications. This paper provides the missing argument. I argue that by virtue of their material properties, technological artifacts are part of the normative order rather than external to it. I illustrate how values can be empirically identified in technology. The reason why value-talk is not trivial or metaphorical is that due to the endurance and longevity of technological artifacts, values embedded in them have long-term implications that surpass their designers and builders. I further argue that taking sides in this debate has real-world implications in the form of moral constraints on the development of technology. (shrink)

In controversies about technology and society, there is no idea more pro vocative than the notion that technical things have political qualities. At issue is the claim that the machines, structures, and systems of modern material culture can be accurately judged not only for their contributions of efficiency and pro-ductivity, not merely for their positive and negative environmental side effects, but also for the ways in which they can embody specific forms of power and authority. Since ideas of this kind (...) have a persistent and troubling presence in discussions about the meaning of technology, they deserve explicit attention... (shrink)

Simple idealized models seem to provide more understanding than opaque, complex, and hyper-realistic models. However, an increasing number of scientists are going in the opposite direction by utilizing opaque machine learning models to make predictions and draw inferences, suggesting that scientists are opting for models that have less potential for understanding. Are scientists trading understanding for some other epistemic or pragmatic good when they choose a machine learning model? Or are the assumptions behind why minimal models provide understanding misguided? In (...) this paper, using the case of deep neural networks, I argue that it is not the complexity or black box nature of a model that limits how much understanding the model provides. Instead, it is a lack of scientific and empirical evidence supporting the link that connects a model to the target phenomenon that primarily prohibits understanding. (shrink)

What responsibility do individuals bear for structural injustice? Iris Marion Young has offered the most fully developed account to date, the Social Connections Model. She argues that we all bear responsibility because we each causally contribute to structural processes that produce injustice. My aim in this article is to motivate and defend an alternative account that improves on Young’s model by addressing five fundamental challenges faced by any such theory. The core idea of what I call the “Role-Ideal Model” is (...) that we are each responsible for structural injustice through and in virtue of our social roles, i.e. our roles as parents, colleagues, employers, citizens, etc., because roles are the site where structure meets agency. In short, the Role-Ideal Model explains how individual action contributes to structural change, justifies demands for action from each particular agent, specifies what kinds of acts should be undertaken, moderates between demanding too much and too little of individual agents, and provides an account of the critical responses appropriate for holding individuals accountable for structural injustice. (shrink)

Visual perception relies on stored information and environmental associations to arrive at a determinate representation of the world. This opens up the disturbing possibility that our visual experiences could themselves be subject to a kind of racial bias, simply in virtue of accurately encoding previously encountered environmental regularities. This possibility raises the following question: what, if anything, is wrong with beliefs grounded upon these prejudicial experiences? They are consistent with a range of epistemic norms, including evidentialist and reliabilist standards for (...) justification. I argue that we will struggle to locate a flaw with these sorts of perceptual beliefs so long as we focus our analysis at the level of the individual and her response to information. We should instead broaden our analysis to include the social structure within which the individual is located. Doing so lets us identify a problem with the way in which unjust social structures in particular “gerrymander” the regularities an individual is exposed to, and by extension the priors their visual system draws on. I argue that in this way, social structures can cap perceptual skill. (shrink)

Call “epistocracy” a political regime in which the experts, those who know best, rule; and call “the epistocratic claim” the assertion that the experts’ superior knowledge or reliability is “a warrant for their having political authority over others.” Most of us oppose epistocracy and think the epistocratic claim is false. But why is it mistaken? Contemporary discussions of this question focus on two answers. According to the first, expertise could, in principle, be a warrant for authority. What bars the successful (...) justification of epistocracy is that the relevant kind of expertise does not exist in politics (either because there are no procedure-independent standards of right or wrong in politics, or because, though such standards exist, no one knows better than anyone else what they require). This skeptical position comes, however, at a significant cost: Without the assumption that some political decisions are better than others, and that some people know better than others what these decisions are, it is difficult to make sense of much of our political practice, including how we criticize politicians and choose among candidates for office. The second answer accepts that there is expertise of the relevant sort in politics. It argues, however, that such expertise does not justify political authority because political justifications are subject to special “acceptability requirements.” Since claims to expertise are normally not acceptable to all qualified (reasonable etc.) points of view, they cannot function as premises in the justification of political authority, and the epistocratic claim fails. Yet as a number of critics have pointed out, this (broadly Rawlsian) strategy faces significant problems: it is at least unclear whether the strategy in fact bars all epistocratic conclusions whether there is any principled way to draw the distinction between qualified and non-qualified points of views on which it depends; and whether principled defenses for it are available and internally consistent. This article outlines a third and previously largely overlooked answer, which resists the epistocratic claim without either denying the existence of expertise in politics or invoking special acceptability requirements for political justifications. The only plausible argument for the epistocratic claim, this article argues, focuses on the compensatory role that the expert’s authority plays in correcting the subject’s relative unreliability or other agential shortcomings. The expert’s authority is thus justified only if the subject, by adopting a policy of obeying the expert’s directives, does not face problems that are very similar to the ones that the expert’s authority was meant to solve in the first place. If, for instance, the subject finds it no easier to reliably identify what the expert’s directives require of him than to reliably assess and act on the reasons with which the expert is meant to help him, then the expert’s directives lack the compensatory value that would justify her authority. But if some widely accepted empirical conjectures about politics in a pluralistic political community are correct, then citizens normally either have no reason to adopt a policy of obeying experts, or the experts with regard to whom they have reason to adopt such a policy differ, so that no expert has the kind of general authority over the polity that we associate with political rule. (We may call this the “non-compensation argument” against epistocracy.) The argument is important both because it helps shed light on the proper relation between authority and expertise in general, and because it shows that we can normally reject the epistocratic claim without adopting either the skeptical or the Rawlsian strategy, thus undercutting whatever support these views derive from the mistaken perception that they are necessary for resisting the threat of epistocracy. Finally, because the anti-epistocratic constraints it introduces apply only to justifications of the subjects’ duty to obey, but not to the existence or activities of political institutions as such, the compensation argument can accommodate our anti-epistocratic intuitions without excluding epistemic considerations from the design of political institutions more generally. (shrink)

One of the most noticeable trends in recent years has been the increasing reliance of public decision-making processes on algorithms, i.e. computer-programmed step-by-step instructions for taking a given set of inputs and producing an output. The question raised by this article is whether the rise of such algorithmic governance creates problems for the moral or political legitimacy of our public decision-making processes. Ignoring common concerns with data protection and privacy, it is argued that algorithmic governance does pose a significant threat (...) to the legitimacy of such processes. Modelling my argument on Estlund’s threat of epistocracy, I call this the ‘threat of algocracy’. The article clarifies the nature of this threat and addresses two possible solutions. It is argued that neither solution is likely to be successful, at least not without risking many other things we value about social decision-making. The result is a somewhat pessimistic conclusion in which we confront the possibility that we are creating decision-making processes that constrain and limit opportunities for human participation. (shrink)

Can the phenomenal character of perceptual experience be altered by the states of one's cognitive system, for example, one's thoughts or beliefs? If one thinks that this can happen then one thinks that there can be cognitive penetration of perceptual experience; otherwise, one thinks that perceptual experience is cognitively impenetrable. I claim that there is one alleged case of cognitive penetration that cannot be explained away by the standard strategies one can typically use to explain away alleged cases. The case (...) is one in which it seems subjects' beliefs about the typical colour of objects affects their colour experience. I propose a two-step mechanism of indirect cognitive penetration that explains how cognitive penetration may occur. I show that there is independent evidence that each step in this process can occur. I suspect that people who are opposed to the idea that perceptual experience is cognitively penetrable will be less opposed to the idea when they come to consider this indirect mechanism and that those who are generally sympathetic to the idea of cognitive penetrability will welcome the elucidation of this plausible mechanism. (shrink)

A story does more than recount events; it recounts events in a way that renders them intelligible, thus conveying not just information but also understanding. We might therefore be tempted to describe narrative as a genre of explanation. When the police invite a suspect to “tell his story,” they are asking him to explain the blood on his shirt or his absence from home on the night of the murder; and whether he is judged to have a “good story” will (...) depend on its adequacy as an explanation. Can we account for the explanatory force of narrative with the models of explanation available in the philosophy of science? Or does narrative convey a different kind of understanding, which requires a different model and perhaps even a term other than ‘explanation’? (shrink)

The question of the relationship of the making of value judgments in a typically ethical sense to the methods and procedures of science has been discussed in the literature at least to that point which e. e. cummings somewhere refers to as “The Mystical Moment of Dullness.” Nevertheless, albeit with some trepidation, I feel that something more may fruitfully be said on the subject.

One theme in complaints from those with marginalized social identities is that they are seen primarily in terms of that identity. Some Black artists, for instance, complain about being seen as Black first and artists second. These individuals can be understood as objecting to a particularly subtle form of morally problematic attention: “relative attentional surplus on the wrong property.” This attentional surplus can coexist with another type of common problematic attention affecting these groups, including attentional deficits; marginalized individuals and groups (...) themselves are routinely insufficiently attended to in virtue of the surplus attention given to their social identity properties. (shrink)

This study investigates a practice that allows workers based in India to work online on projects for corporations in the United States, representing a new mode of labor integration. In the absence of direct bureaucratic control across continents, the question arises how this rapidly growing labor practice is organized. The riddle of organizational governance is solved through an analysis of software programming schemes, which are presented as the key to organizing globally dispersed labor through data servers. This labor integration through (...) programming code is distinguished from two other systems of organization—bureaucracy and the market—while bringing out the salient features of each system in terms of its ruling principle: bureaucracy (legalrational), the market (price), and algocracy (programming or algorithm). The logic of algocratic systems is explored methodically to analyze global work. (shrink)