There are many historical concerns about freedom that have come to be deemphasized in the free will literature itself—for instance, worries around the tyranny of government or the alienation of capitalism. It is hard to see how the current free will literature respects these, or indeed how they could even find expression. This paper seeks to show how these and other concerns can be reintegrated into the debate by appealing to a levels ontology. Recently, Christian List and others have considered (...) how the notion of levels could be relevant to the free will debate. Invariably, however, the focus is on the significance of facts at lower levels. The threats come from below, from fundamental physics or neuroscience. Here, I aim to show how we can frame many interesting concerns about free will in terms of threats from above. After arguing that determination from above is no less threatening, I catalogue such concerns that might constitute threats to our freedom. Doing this not only allows us to show how these concerns relate to those standardly discussed, but it pushes us to expand our conception of freedom. (shrink)

Libertarian paternalism is a weak form of paternalism that recommends nudges rather than bans, restrictions, or other strong interventions. Nudges influence people’s choice by modifying contextual factors (the “choice architecture”). This paper explores the possibility of an epistemic analogue of libertarian paternalism. What I call “epistemic libertarian paternalism” is a weak form of epistemic paternalism that recommends “epistemic nudges” rather than stronger paternalistic interventions. Epistemic nudges influence people’s beliefs and judgments by modifying contextual factors (the “epistemic choice architecture”). The main (...) aim of this paper is to defend epistemic libertarian paternalism from the “irrationality problem”, which I take to be the most urgent problem for epistemic libertarian paternalism; given how epistemic nudges work (i.e. they typically co-opt psychological biases), nudged beliefs are irrational. In response to the irrationality problem, I admit that nudged beliefs are often (not always, though) irrational, but insist that there are conditions in which epistemic nudging can be justifiable nonetheless. I will propose two conditions that are jointly sufficient for justifiable epistemic nudging: “Veridicality Condition” (which says that nudged beliefs are more likely to be true than non-nudged beliefs) and “Not-More-Irrationality Condition” (which says that nudged beliefs are not more likely to be irrational than non-nudged beliefs). (shrink)

Behavioural techniques or ‘nudges’ can be used for various purposes. In this paper, we shift the focus from government nudges to nudges used by for-profit market agents. We argue that potential worries about nudges circumventing the deliberative capacities or diminishing the control of targeted agents are greater when it comes to market nudges, given that these (1) are not constrained by the principles that regulate government nudges (mildness, sensitivity to people’s interests and public justifiability) and (2) are often ‘stacked’ – (...) they come in great numbers that overwhelm agents. In addition, we respond to possible objections and derive several policy suggestions. (shrink)

Nudging works through dispositions to decide with specific heuristics, and has three component parts. A nudge is a feature of an environment that enables such a disposition; a person is nudged when such a disposition is triggered; and a person performs a nudged action when such a disposition manifests in action. This analysis clarifies an autonomy-based worry about nudging as used in public policy or for private profit: that a person’s ability to reason well is undermined when she is nudged. (...) Reasoning well is a component of self-guidance, which is an aspect of autonomy and so something there is reason to promote, preserve, or respect. However, a person can reason well when she is nudged: Many of these heuristics are good rules to reason with, and she can be flexible with respect to them when she takes there to be a better way to reason. Along the way, this paper uncovers several design specifications for responsible nudging, and discusses open empirical questions. However, nudging’s being compatible with reasoning well crystallizes a distinct worry about manipulation: that nudge designers can rely on nudged people guiding themselves toward the designers’ own ends. Manipulation of this sort exploits one aspect of autonomy (namely, self-guidance) to undermine autonomy in other respects. (shrink)

A central reason to worry that managers should not use nudges to influence employees is that doing so fails to treat employees as _rational_ and/or _autonomous_ (RA). Recent nudge defenders have marshaled a powerful line of response against this worry: in general, nudges treat us as the kind of RA agents we are, because nudges are apt to enhance our limited capacities for RA agency by improving our decision-making environments. Applied to managerial nudges, this would mean that when managers nudge (...) their employees, they generally bolster their employees’ limited RA agency and, thus, treat employees as the kinds of RA agents they are. My aim is to vindicate a qualified version of the initial worry from the nudge-defender response and, as a result, provide a clearer, more plausible framework for evaluating managerial nudges than what nudge critics have previously given. I do this, first, by showing how nudge defenders rely on equivocation between two different senses of “treating someone as RA.” The _value-preserving_ notion that supports the nudge-defender prescription to protect and enhance RA capacities is different from the _authority-recognizing_ notion that underwrites the initial worry about nudging. Second, I argue that the authority-recognizing notion of treating someone as RA implies that managerial nudges treat employees as RA just when the nudges are compatible with relating to employees as equals. Third, I explain how, to determine when managerial nudges are compatible with relating to employees as equals, we need to consider how employees surrender aspects of their equal, agency-grounded authority to managers. (shrink)

This paper focuses on the potential of “equitech”—AI technology that improves equity. Recently, interventions have been developed to reduce the harm of implicit bias, the automatic form of stereotype or prejudice that contributes to injustice. However, these interventions—some of which are assisted by AI-related technology—have significant limitations, including unintended negative consequences and general inefficacy. To overcome these limitations, we propose a two-dimensional framework to assess current AI-assisted interventions and explore promising new ones. We begin by using the case of human (...) resource recruitment as a focal point to show that existing approaches have exploited only a subset of the available solution space. We then demonstrate how our framework facilitates the discovery of new approaches. The first dimension of this framework helps us systematically consider the analytic information, intervention implementation, and modes of human-machine interaction made available by advancements in AI-related technology. The second dimension enables the identification and incorporation of insights from recent research on implicit bias intervention. We argue that a design strategy that combines complementary interventions can further enhance the effectiveness of interventions by targeting the various interacting cognitive systems that underlie implicit bias. We end with a discussion of how our cognitive interventions framework can have positive downstream effects for structural problems. (shrink)

In discussions of nudging, transparency is often taken to be important; it is often suggested that a significant moral consideration to take into account when nudging is whether the nudge is transparent. Another consideration taken to be relevant is whether the nudge is easy to resist. Sometimes, these two considerations are taken to be importantly related: if we have reason to make nudges easy to resist, then we have reason to make them transparent, insofar as a nudge’s transparency is relevant (...) to whether the nudge is easy to resist. In this paper, we critically scrutinize this view. First, we draw out the purported connections between transparency and resistibility, combining them into what we call the Resistibility-to-Transparency Argument (RTA). Then, we do some conceptual groundwork elaborating on different forms of transparency mentioned in the debate. With this in place, we argue that in order to be plausible, the RTA must appeal to certain forms of transparency; those on which the transparency of a nudge depends on the nudgee’s ability to become aware of the relevant facts about the nudge. We then argue against the common assumption that in order for a nudge to be easy to resist for an individual, that individual needs to have the capacity to easily become aware of the relevant nudge facts. We conclude that the connection between the easy resistibility of a nudge and its transparency is, at best, a weak one. (shrink)

Nudges are a way to steer people’s behavior through changes in how choices are presented. Nudge research has been incorporated into public policy in many countries, and nudge research, thus, has the potential to directly influence societies and individuals. As a result, research ethics for nudge research is needed to ensure that nudges developed are not instances of unethical manipulation of people. In this paper, I argue that two types of slippery slopes from ethically fine nudges to ethically problematic ones (...) can take place in nudge research. The conceptual slippery slope follows from (1) the broad way of defining nudges, (2) the multitude of different ways of understanding manipulation of people, (3) many manipulation definitions implying that some nudges are manipulative, and (4) some forms of manipulation of people being prima facie wrong. The empirical slippery slope rests, first, on the possibility of small steps from ethically acceptable to wrong nudges, and second, on the cognitive limits typical to human beings (including nudge researchers). Both slippery slope arguments imply the need to of nudge researchers to create so-called firebreaks. The firebreaks can take the form of ethical justifications. To avoid slippery slopes, certain types of arguments should be excluded from these justifications. (shrink)

Much discussion of the ethics of algorithms has focused on harms to autonomy—especially harms stemming from manipulation. Nonetheless, although manipulation can often be harmful, we suggest that in certain contexts it may not impair autonomy. To fully assess the impact of algorithm design on autonomy, we argue for a need to move beyond a focus on manipulation towards a multidimensional account of autonomy itself. Drawing on the autonomy literature and recent data ethics, we propose a novel account which takes autonomy (...) to supervene on three distinct but related elements: agency, authenticity and individual control over decision-making. By distinguishing autonomy from control, the account can explain the variable effects of manipulation on user autonomy within algorithm-driven systems. In particular, it can explain why improving user control may improve autonomy in some contexts, while in other contexts—such as in some kinds of newsfeeds—certain algorithm designs that instead _reduce_ user control may nevertheless improve autonomy. As a result, the account can accommodate the sometimes convoluted interplay between control, autonomy, manipulation, and commercial versus prosocial design goals. (shrink)

Nudges are means to influence the will formation of people to make specific choices more likely. My focus is on nudges that are supposed to improve the health condition of individuals and populations over and above the direct prevention of disease. I point out epistemic and moral problems with these types of nudges, which lead to my conclusion that health-enhancing nudges fail. They fail because we cannot know which choices enhance individual health—properly understood in a holistic way—and because health-enhancing nudges (...) are often themselves bad for our health. They can be bad for our health because they assume inferior agency in their targets and accordingly regularly lead to appropriate resentment and anger—strong emotions which go along with an increased risk of health impairments. Briefly, health-enhancing nudges fail because they are based on persistent ignorance and on a presumptuous attitude. (shrink)

This paper explores the extent to which behavioral public policies can be both efficient and democratic by reflecting on the conditions under which individuals could rationally consent to them. Consent refers to a moral requirement that a behavioral public policy should respect what I call a person’s value autonomy and conception of the good. Behavioral public policies can take many forms. Based on a social choice framework, I argue that fully paternalistic and prudential behavioral public policies are unlikely to trigger (...) a hypothetical form of rational consent. Non-fully paternalistic behavioral public policies with partially non-prudential motivations are less problematic in this perspective. In any case, a public deliberative stage preceding the implementation of policies seems to be the best democratic way to justify them. (shrink)