Anna Hartford

The focus of debates about conversational AIs (CAIs) has largely been on social and ethical concerns that arise when we speak to machines. What is gained and what is lost when we replace our human interlocutors—including our human therapists— with AIs? Here, we focus instead on a distinct and growing phenomenon: letting machines speak for us. What is at stake when we replace our own efforts at interpersonal engagement with CAIs? The purpose of these technologies is, in part, to remove effort. But effort also has enormous value, and in some cases even intrinsic value. This is true in many realms, but especially in interpersonal relationships. To make an effort for someone, irrespective of what that effort amounts to, often conveys value and meaning in itself. We elaborate on the meaning, worth and significance that may be lost when we relinquish effort in our interpersonal engagements, and also on the opportunities for self-understanding and growth that we may forsake.

Despite growing understanding of the addictive qualities of the internet, and rising concerns about the effects of excessive internet use on personal wellbeing and mental health, the corresponding ethical debate is still in its infancy, and many of the relevant philosophical and conceptual frameworks are underdeveloped. Our goal in this chapter is to explore some of this evolving terrain. While there are unique ethical considerations that pertain to the formalisation of a disorder related to excessive internet use, our ethical concerns (and indeed our mental health concerns) about whether certain technologies undermine wellbeing can and should be far broader than the debates about the appropriateness of particular diagnostic categories. In this chapter we introduce some of these wider debates with regards to persuasive digital technologies— particularly those which aim to maximise use, or even to encourage compulsive engagement—as well as the difficulty in articulating the harms involved in excessive internet use, especially where such use has not led to functional impairment. Following these conversations we briefly consider some more practical ethical implications, including regulation of certain design features, concerns about growing socioeconomic inequality in online services, and whether there should be a “right to disconnect.”

We routinely take diminished capacity as diminishing moral responsibility (as in the case of immaturity, senility, or particular mental impairments). The prospect of enhanced capacity therefore poses immediate questions with regard to moral responsibility. Of particular interest are those capacities that might allow us to better avoid serious harms or wrongdoing. We can consider questions of responsibility with regards to enhancement at various removes. In the first instance: where such (safe and effective) interventions exist, do we have an obligation to undergo such enhancement? Secondly: once enhanced, would the ambit of our responsibility therefore increase? Some philosophers have argued that enhanced capacities potentially generate “hyperresponsibility.” Hyperresponsible people would be held to a different and higher moral standard than those of us with more ordinary human capacities, and are liable to be more blameworthy for wrongdoing than ordinary agents. This chapter discusses the implications of enhancement for three central views of responsibility, namely: capacity-based, control-based and revelation-based views. Debates around moral responsibility have primarily concerned diminished capacities; as such the prospect of enhancement introduces new terrain—and potentially new fault-lines and complexities—in which to interrogate our theoretical conceptions of the foundations (and limits) of moral responsibility.

Recent years have seen growing public concern about the effects of persuasive digital technologies on public mental health and well-being. As the draws on our attention reach such staggering scales and as our ability to focus our attention on our own considered ends erodes ever further, the need to understand and articulate what is at stake has become pressing. In this ethical viewpoint, we explore the concept of attentional harms and emphasize their potential seriousness. We further argue that the acknowledgment of these harms has relevance for evolving debates on digital inequalities. An underdiscussed aspect of web-based inequality concerns the persuasions, and even the manipulations, that help to generate sustained attentional loss. These inequalities are poised to grow, and as they do, so will concerns about justice with regard to the psychological and self-regulatory burdens of web-based participation for different internet users. In line with calls for multidimensional approaches to digital inequalities, it is important to recognize these potential harms as well as to empower internet users against them even while expanding high-quality access.

An area of consensus in debates about culpability for ignorance concerns the importance of an agent’s epistemic situation, and the information available to them, in determining what they ought to know. On this understanding, given the excesses of our present epistemic situation, we are more culpable for our morally-relevant ignorance than ever. This verdict often seems appropriate at the level of individual cases, but I argue that it is over-demanding when considered at large. On the other hand, when we describe an obligation to know that avoids over-demandingness at large, it fails to be sufficiently demanding in individual cases. The first half of this paper is dedicated to setting up this dilemma. In the second half, I show that it cannot be easily escaped. Finally, I suggest that this dilemma impedes our ability to morally appraise one another’s ignorance, and even our own.

Difficulty is often treated as blame-mitigating, and even exculpating. But on some occasions difficulty seems to have little or no bearing on our assessments of moral responsibility, and can even exacerbate it. In this paper, I argue that the relevance (and irrelevance) of difficulty with regard to assessments of moral responsibility is best understood via Quality of Will accounts. I look at various ways of characterising difficulty – including via sacrifice, effort, skill and ‘trying’ – and set out to demonstrate that these factors are only blame-mitigating where, and to the extent that, they complicate ascriptions of insufficient concern. Matters become more complex, however, when we turn to difficult circumstances that seem to generate such objectionable attitudes. This is arguably the case with epistemic difficulty and certain instances of moral ignorance. Here I argue that certain difficult circumstances diminish the sense in which false moral beliefs are genuinely revelatory of the agents who hold them. In particular, I draw on the distinction between difficulty that generates objectionable attitudes, and objectionable attitudes that generate difficulty. I argue that the former, but not the latter, can plausibly be viewed as blame mitigating, and that this would apply to (limited) cases of moral ignorance.

Within debates concerning responsibility for ignorance the distinction between moral and factual ignorance is often treated as crucial. Many prominent accounts hold that while factual ignorance routinely exculpates, moral ignorance never does so. The view that there is an in-principle distinction between moral and factual ignorance has been referred to as the “Asymmetry Thesis.” This view stands in opposition to the “Parity Thesis,” which holds that moral and factual ignorance are in-principle similar. The Parity Thesis has been closely aligned with volitionist accounts of moral responsibility, whereas the Asymmetry Thesis has been closely aligned with Quality of Will accounts. Two central questions are at work here: how ignorance excuses, and whether it excuses in the same way for both moral and factual ignorance. I will argue that these questions have often been confused in the present debate, and once we have distinguished more clearly between them, it seems that Quality of Will accounts are compatible with the Parity Thesis. And more generally: that the distinction between moral and factual ignorance is far less important in debates about responsibility for ignorance than it has often appeared.

I seek to re-approach the longstanding debate concerning the moral relevance of physical distance by emphasising the important distinction between evaluations of wrongdoing and evaluations of blameworthiness. Drawing in particular on Quality of Will accounts of blameworthiness, I argue that proximity can make an important difference to what qualifies as sufficient moral concern between strangers, and therefore to evaluations of blameworthiness for failures to assist. This implies that even if two individuals (one distant, one proximate) commit an equivalent wrong in ignoring preventable suffering, it can still be the case that the proximate individual is more blameworthy insofar as their disregard expresses a more reprehensible attitude. I argue that emphasising this distinction allows us to make sense of a range of conflicting intuitions with regards to the effects of physical distance on our obligations to assist.

The idea that conscious control, or more specifically akratic wrongdoing, is a necessary condition for blameworthiness has durable appeal. This position has been explicitly championed by volitionist philosophers, and its tacit influence is broadly felt. Many responses have been offered to the akrasia requirement espoused by volitionists. These responses often take the form of counterexamples involving blameworthy ignorance: i.e., cases where an agent didn’t act akratically, but where they nevertheless seem blameworthy. These counterexamples have generally led to an impasse in the debate, with volitionists maintaining that the ignorant agents are blameless. In this paper, I explore a different sort of counterexample: I consider agents who have acted akratically, but whose very conscious awareness of their wrongdoing complicates their blameworthiness. I call these cases of “complex akrasia,” and I suggest that they are a familiar aspect of moral life. I interpret these cases as supporting non-volitionist accounts, and particularly Quality of Will accounts.