Agent-Centered Deontological Theories

The field of AI safety aims to prevent increasingly capable artificially intelligent systems from causing humans harm. Research on moral alignment is widely thought to offer a promising safety strategy: if we can equip AI systems with appropriate ethical rules, according to this line of thought, they'll be unlikely to disempower, destroy or otherwise seriously harm us. Deontological morality looks like a particularly attractive candidate for an alignment target, given its popularity, relative technical tractability and commitment to harm-avoidance principles. I (...) argue that the connection between moral alignment and safe behavior is more tenuous than many have hoped. In general, AI systems can possess either of these properties in the absence of the other, and we should favor safety when the two conflict. In particular, advanced AI systems governed by standard versions of deontology need not be especially safe. (shrink)

When, if ever, do we wrong ourselves? The Self-Other Symmetric answer is: when we do to ourselves what would wrong a consenting other. The standard objection, which has gone unchallenged for decades, is that Symmetry seems to imply that we wrong ourselves in too many cases—where rights are unwaivable, or “self-consent” is lacking. We argue that Symmetry not only survives these would-be counterexamples; it explains and unifies them. The key to Symmetry is not, as critics have supposed, the bizarre claim (...) that we must literally give ourselves consent if we are to avoid wronging ourselves. Instead, it is that we authorize ourselves simply by making decisions, just as we can authorize others by making decisions jointly. (shrink)

We argue that supererogation cannot be understood just in terms of reasons for action. In addition to reasons, a theory of supererogation must include prerogatives, which can make an action permissible without counting in favor of doing it.

Fiona Woollard argues that when one is personally involved in an emergency, one has a moral requirement to make substantial sacrifices to aid others that one would not otherwise have. She holds that there are three ways in which one could be personally involved in an emergency: by being physically proximate to the victims of the emergency; by being the only person who can help the victims; or by having a personal encounter with the victims. Each of these factors is (...) claimed to be defeasibly sufficient to ground personal involvement, and thus a requirement of substantial sacrifice to aid. Woollard defends this view on the basis of a number of cases. We show that Woollard’s cases contain various confounding factors. In view of the more precisely drawn cases offered here, it is clear that neither proximity nor uniqueness nor personal encounter is intuitively defeasibly sufficient in the way Woollard claims. (shrink)

If we were required to sacrifice our own interests whenever doing so was best overall, or prohibited from doing so unless it was optimal, then we would be mere sites for the realisation of value. Our interests, not ourselves, would wholly determine what we ought to do. We are not mere sites for the realisation of value — instead we, ourselves, matter unconditionally. So we have options to act suboptimally. These options have limits, grounded in the very same considerations. Though (...) not merely such sites, you and I are also sites for the realisation of value, and our interests (and ourselves) must therefore sometimes determine what others ought to do, in particular requiring them to bear reasonable costs for our sake. Likewise, just as my moral status grounds a requirement that others show me appropriate respect, so must I do to myself. (shrink)

Some sacrifices—like giving a kidney or heroically dashing into a burning building—are supererogatory: they are good deeds beyond the call of duty. But if such deeds are really so good, philosophers ask, why shouldn’t morality just require them? The standard answer is that morality recognizes a special role for the pursuit of self-interest, so that everyone may treat themselves as if they were uniquely important. This idea, however, cannot be reconciled with the compelling picture of morality as impartial—the view that (...) we are each anyone’s equal. I propose an alternative Self-Other Symmetric account of our moral freedom: the limits on what morality may demand of us are set by the duties we owe to ourselves. I begin with a defense of the Self-Other Symmetry: the idea that we owe the same duties to ourselves—and have the same rights against ourselves—as any relevantly similar other. Because we are consenting parties to our own actions, I argue, our rights against ourselves do not function like the rights of unwilling others. Instead, they play a permissive function, allowing us to block the demand to give up what is ours. I conclude by uniting, aggravating, and trying to solve some paradoxes of supererogatory permissions, guided by the idea that morality cannot be reduced to a ranking of options from best-to-worst. Rights against oneself are an irreducible second dimension, one that we need if we are to unify rights and supererogation into an impartial moral vision. (shrink)

Many of us give to charities that are close to our hearts rather than those that would use our gifts to do more good, impartially considered. Is such partiality to charities acceptable? I argue that if partiality to particular people is justified, we can go SOME distance toward justifying partiality to particular charities. Even so, partiality to charities is justified in fewer cases than most people seem to believe.

Many of us think we have agent-centred options to act suboptimally. Some of these involve favouring our own interests. Others involve sacrificing them. In this paper, I explore three different ways to accommodate agent-centred options in a criterion of objective permissibility. I argue against satisficing and rational pluralism, and in favour of a principle built around sensitivity to personal cost.

Theorists have argued that under certain background conditions the commercial, for-profit corporation might bear responsibility to act to advance justice. However, other agents too may be responsible to take remedial action, especially when the state defaults. This raises the question of the sequence in which the agents should act. I develop a framework that offers guidance in determining when the corporation ought to intervene to advance justice. The existing literature typically identifies responsibility-bearers solely by their capacity to remedy an unjust (...) situation which I believe to be too simplistic. I introduce two additional grounds for identifying responsibility-bearers, a role-based account and participation-based account and show that this pluralist approach delivers a better account of who bears responsibility to act and when to discharge this responsibility. (shrink)

This book explores a universal question of human social order: Under what circumstances and to what extent is the individual to be held morally responsible for collective events? This question reaches far beyond the intentions and actions of a particular business enterprise, state or a similar large-scale collective. The philosopher Wolfgang Sohst (Berlin, Germany) investigates the subject with unprecedented thoroughness, covering the whole range of contemporary discussion on this subject. He provides a detailed analysis of the functions of individual members (...) in such a collective, the structural prerequisites for them to be held responsible for acts which they have not directly committed themselves and the transmission of responsibility even to successor generations of the perpetrators collective. -/- Table of contents: -/- Introduction 1 Actors and Moral Action 1.1 On the concepts of actors and action 1.2 The simultaneous emergence of actor and action 1.3 The difference between a unit of event and a unit of action 1.4 The difference between legal and moral responsible action -/- 2 The Continuum Between Individual and Corporate Actor 2.1 The relationship between a single human actor and a corporate actor 2.1.1 The primary responsibility of the individual actor and the ontological status of the collective 2.1.2 Additional arguments for putting collective responsibility onto the individual actor 2.2 The levels within structural consolidation 2.3 A different schematic view: Community, Society, State 2.4 Direct vs. organized sociality 2.5 Conflicts of application in assigning collective responsibility -/- 3 Individual and Collective Actions 3.1 A better form of social reductionism 3.2 Corporate bodies as a bundle of agency relationships 3.3 The horizon of view for collective social phenomena -/- 4 The Corporate Entity as a Moral Subject 4.1 Are corporate actors also morally responsible? 4.2 Corporate bodies as norm subjects -/- 5 Possible criteria for the moral qualification of collective action 5.1 Membership in a group 5.2 Success of an action 5.3 Shared intentionality or purpose, common interests and common consciousness 5.4 Social relationships between actors as a condition of collective action 5.5 Subjective and factual feelings of collective responsibility 5.6 The community of shared values 5.7 Social identity 5.8 Origin and ethnic belonging 5.9 No equality in injustice -/- 6 Norm-based and Purpose-oriented Organization 6.1 No collective responsibility without inner organization 6.2 Collective shame as an indication of collective responsibility 6.3 Individual responsibility for collective norms 6.4 The collective organization as an independent unit of purpose for the collective 6.4.1 The actualistic perspective 6.4.2 Structural persistence 6.4.3 Possible counter-examples of structural persistence 6.5 The other side of collective responsibility 6.6 Interim Result -/- 7 Moral Responsibility of the Individual from an International Perspective -/- 8 Social Norms and Our Responsibility for their Fulfillment 8.1 Norm dimensions 8.2 The overarching importance of norm ranking in assigning collective moral responsibility 8.3 Private and public norms 8.4 Subjective ‘ought’ and personal responsibility 8.5 The obligation to morally acceptable and coherent behavior -/- 9 The Difference between Culpability and Responsibility 10 The Temporal Horizon of Collective Moral Responsibility Index Bibliography. (shrink)

Effective altruists recommend that we give large sums to charity, but by far their more central message is that we give effectively, i.e., to whatever charities would do the most good per dollar donated. In this paper, I’ll assume that it’s not wrong not to give bigger, but will explore to what extent it may well nonetheless be wrong not to give better. The main claim I’ll argue for here is that in many cases it would be wrong of you (...) to give a sum of money to charities that do less good than others you could have given to instead, even if it would not have been wrong of you not to give the money to any charity at all. I assume that all the charities under discussion here do positive good overall, do not cause harm, do not infringe rights, etc. What makes my main claim here particularly interesting is that it is inconsistent with what appears to be a fairly common assumption in the ethics of giving, according to which if it is not wrong of you to keep some sum of money for yourself, then it is likewise not wrong of you to donate it to any particular charity you choose. Roughly: if it’s up to you whether to donate the money, it’s also up to you where to donate the money. I challenge this common assumption. (shrink)

Is there a reason to prevent deontological wrongdoing—an action that is wrong due to the violation of a decisive deontological constraint? This question is perplexing. On the one hand, the intuitive response seems to be positive, both when the question is considered in the abstract and when it is considered with regard to paradigmatic cases of deontological wrongdoing such as Bridge and Transplant. On the other hand, common theoretical accounts of deontological wrongdoing do not entail this answer, since not preventing (...) wrongdoing does not necessarily amount to doing harm or intending harm, for example. The puzzle is reinforced due to the fact that the intuitive response to other cases seems to be different, namely that there is no reason in favour of preventing deontological wrongdoing. This question is thus interesting in itself. It might also shed light on additional questions such as the “paradox of deontology” and the appropriate response to wrongful actions more generally. Yet, despite its importance, this question is typically overlooked. The paper explores this question. (shrink)