Stefano Calboli

In recent years, there has been growing enthusiasm for leveraging artificial intelligence (AI) to enhance the effectiveness of nudges. However, this enthusiasm has been tempered by significant ethical concerns, one of which relates to the political opacity of the psychological mechanisms underlying AI-driven nudges. Political transparency is widely considered essential for enabling meaningful public scrutiny and the political opacity often associated with AI-driven nudges appears to undermine citizens’ ability to critically assess them. This concern introduces a potential ethical challenge that sets AI-driven nudges apart from traditional ones, which are typically assumed to be transparent about the psychological mechanisms they are based upon. This paper seeks to critically assess this concern and argue that it is ultimately unwarranted. First, I will illustrate that the empirical basis for this worry is flawed, as traditional nudges can also exhibit political opacity in relation to their psychological mechanisms. Second, I will argue that citizens are fully capable of engaging in meaningful public scrutiny even without complete knowledge of the psychological mechanisms underlying nudges. I contend that denying public scrutiny on the grounds of incomplete knowledge is paternalistic, infantilizing, and, paradoxically, undermines the very principles that justify the practice of public scrutiny itself.

What role can artificial intelligence (AI) play in enhancing public policy nudges and the extent to which these help people achieve their own goals? Can it help mitigate or even overcome the challenges that nudgers face in this respect? This paper discusses how AI-enhanced personalization can help make nudges more means paternalistic and thus more respectful of people’s ends. We explore the potential added value of AI by analyzing to what extent it can, (1) help identify individual preferences and (2) tailor different nudging techniques to different people based on variations in their susceptibility to those techniques. However, we also argue that the successes booked in this respect in the for-profit sector cannot simply be replicated in public policy. While AI can bring benefits to means paternalist public policy nudging, it also has predictable downsides (lower effectiveness compared to the private sector) and risks (graver consequences compared to the private sector). We discuss the practical implications of all this and propose novel strategies that both consumers and regulators can employ to respond to private AI use in nudging with the aim of safeguarding people’s autonomy and agency.

This article discusses the contributions that information-access-facilitator technologies promise to deliver in implementing ethical safeguards in the use of nudging techniques. It begins by identifying two broad issues within the debate on the ethics of nudging: first, the risk of misalignment between will and action that nudges can impose, and second, citizens' tendency to be oblivious to normatively relevant information regarding nudges. The subsequent analysis focuses on the safeguards intended to mitigate these two issues. Proceeding with the main argument, the article contends that leveraging technological potential can enhance the effectiveness of safeguards. Technologies meant to facilitate access to information, such as apps, Internet of Things products, wearable technologies, virtual assistants, and augmented reality applications, can be used as friction minimizers for obtaining information, saliency enhancers, enablers of information retrieval, and proactive measures to mitigate misjudgments, and can be characterized by a polymorphic warning design, resulting in nudging interventions that optimally fulfill ethical demands.

_Abstract_: Nudges have proven to be effective tools for steering citizens toward desirable behaviors and make valuable additions to any policy-maker’s toolbox. Disappointingly, however, there are no mechanistic explanations for how nudges work, leaving policy-makers unable to explain what happens when they are implemented. This paper identifies some neglected ethical implications of the resulting citizens lack of awareness of such mechanisms. We first examine mechanistic explanations in relation to citizens’ understanding on how they work. Then, we look at mechanistic explanations in light of the suggestion advanced by some ethicists that nudges be considered ethically acceptable in modern liberal democracies provided the explicit transparency of the nudges employed. _Keywords_: Nudge; Ethics of Nudging; Mechanistic Evidence; In-principle Transparency; Explicit Transparency _Spiegazioni meccanicistiche e l’etica dei _nudges _Riassunto_: I _nudge_ si sono rivelati strumenti di _policy_ efficaci nello "spingere gentilmente" i cittadini verso comportamenti considerati desiderabili. Per questa ragione i nudge sono considerati validi strumenti della cassetta degli attrezzi di un _policymaker_. Tuttavia, è in qualche modo deludente la mancanza di spiegazioni meccanicistiche dei nudge, dalla quale risulta l'incapacità dei _policymaker_ di spiegare quello che succede quando i nudge vengono effettivamente implementati. Questo articolo identifica alcune implicazioni etiche finora trascurate connesse alla inconsapevolezza da parte dei cittadini circa i meccanismi che governano i _nudge_. Da una parte, le nostre considerazioni metteranno in relazione le spiegazioni meccanicistiche con la comprensione dei _nudge_ da parte dei cittadini. Dall'altra parte, metteremo in relazione le spiegazioni meccanicistiche con il concetto di trasparenza esplicita, ossia la richiesta avanzata da alcuni eticisti al fine di considerare i _nudge_ eticamente accettabili nelle moderne democrazie liberali. _Parole chiave_: Nudge; Etica dei nudge; Spiegazioni meccanicistiche; Trasparenza implicita; Trasparenza esplicita.

SARS-CoV-2 vaccines are indispensable allies in the fight against COVID-19. Behavioral and cognitive scientists have argued for taking advantage of insights from their fields of investigations in shaping anti-COVID policies. B&C scientists extensively discussed the methodological and practical issues that arise in translating B&C research results into policy interventions aimed to boost vaccination, Nevertheless, the same cannot be said for the ethical aspects. In the present work, we discuss the ethics of nudging vaccination in light of the “alien control” objection, a kind of control whereby an agent has the power to influence the choice of another agent, and the latter cannot control such power. We proceed as follows. We begin by presenting several cases of mandatory policies avaiable to policymakers to boost the vaccine rate. Next, we point out that insights from B&C sciences can inform mandatory policies and open up ways to enhance them. In section 3, we focus on untraditional policy tools, viz. nudges. In the second part of the article, we introduce the debate on the ethics of nudges, focusing on the “alien control” objection and its relevance for the citizens’ freedom and autonomy. We therefore discuss the original point of the paper. We argue that, referring to nudging vaccination, a somehow neglected ethical issue posed by nudges emerges. Our point is that nudges do not exert exclusively an influence over a targeted behavior. Indeed, nudges impose as well a political influence by which citizens are treated as means to succeed in achieving a peculiar, and potentially controversial, political aim. We argue that, due to what we propose to call the “political multi-justifiability" of nudges, their political influence cannot be detected by citizens. This leads to the threat of alien control. In the last section, we draw some tentative conclusions, suggest a possible solution and its practical advantages.