Rights in the Balance

Professor Walen’s book rejects the familiar argument of “double effect,” namely the doctrine that an action that knowingly causes the death of another person cannot be justified merely by its good consequences but only by its good intentions. Professor Walen offers a rival argument. He proposes that we rethink the killing of non-combatants in war on the basis of a theory of “the mechanics of claims” so that the intentional killing of civilians may be occasionally permissible. Such targeting of civilians may be justified, according to the book’s argument, by the aim of eliminating the threat that these civilians may pose—innocently or not—to other persons. In these circumstances, it will not only be permissible, but it would also be a matter of right to kill civilians, which would be derived from a balancing of “claims.” The argument is impressively made but is ultimately unconvincing. All the decisive questions appear to be matters of a balance of “goodness.” The “mechanics of claims” organizes a structure of welfare values that ultimately work as a proxy for act-utilitarianism. As a result, the argument is open to well-known objections regarding justice and the separateness of persons.

Algorithmic Ethics: Formalization and Verification of Autonomous Vehicle Obligations

In this article, we develop a formal framework for automatic reasoning about the obligations of autonomous cyber-physical systems, including their social and ethical obligations. Obligations, permissions, and prohibitions are distinct from a system's mission, and are a necessary part of specifying advanced, adaptive AI-equipped systems. They need a dedicated deontic logic of obligations to formalize them. Most existing deontic logics lack corresponding algorithms and system models that permit automatic verification. We demonstrate how a particular deontic logic, Dominance Act Utilitarianism (DAU) [23], is a suitable starting point for formalizing the obligations of autonomous systems like self-driving cars. We demonstrate its usefulness by formalizing a subset of Responsibility-Sensitive Safety (RSS) in DAU; RSS is an industrial proposal for how self-driving cars should and should not behave in traffic. We show that certain logical consequences of RSS are undesirable, indicating a need to further refine the proposal. We also demonstrate how obligations can change over time, which is necessary for long-term autonomy. We then demonstrate a model-checking algorithm for DAU formulas on weighted transition systems and illustrate it by model-checking obligations of a self-driving car controller from the literature.

Consequentialism

This chapter explains the consequentialist approach to ethical analysis. It distinguishes act utilitarianism and rule utilitarianism. It also considers different possibilities as to which outcomes should be considered relevant for consequentialists. It considers a number of challenges and objections to consequentialist ethics.

Williams on the problem of moral integrity

In this paper, we have tried to point out the importance of the problem of moral integrity in ethical theory. The best way to make an introduction to ethical weighing when it comes to the problem of moral integrity is to analyze the dispute initiated by Bernard Williams. Namely, this is a critique of the act utilitarianism, whose essential weight is precisely on the topic of moral integrity. Williams conceived his objection as saying that there was no place for the value of moral integrity within the act utilitarianism. The treatment of moral integrity is the point of radical disagreement between utilitarianism and deontological ethics. In this way, deciding between utilitarianism and an alternative ethical position, we are actually deciding in favor or against the affirmation of the values of moral integrity. This is a very significant decision when it comes to ethical position. This paper is part of a broader topic on the place of moral integrity in ethical theory, where we have argued that utilitarianism is not the optimal ethical position, precisely because it does not affirm the value of the moral integrity of the individual. This paper is a preparation of such an attitude and has a more modest ambition - it deals with the re-examination of the perception of moral integrity within a utilitarian ethical position.

Whose morality? Which rationality? Challenging artificial intelligence as a remedy for the lack of moral enhancement

Moral implications of the decision-making process based on algorithms require special attention within the field of machine ethics. Specifically, research focuses on clarifying why even if one assumes the existence of well-working ethical intelligent agents in epistemic terms, it does not necessarily mean that they meet the requirements of autonomous moral agents, such as human beings. For the purposes of exemplifying some of the difficulties in arguing for implicit and explicit ethical agents in Moor’s sense, three first-order normative theories in the field of machine ethics are put to test. Those are Powers’ prospect for a Kantian machine, Anderson and Anderson’s reinterpretation of act utilitarianism and Howard and Muntean’s prospect for a moral machine based on a virtue ethical approach. By comparing and contrasting the three first-order normative theories, and by clarifying the gist of the differences between the processes of calculation and moral estimation, the possibility for building what—one might call strong “moral” AI scenarios—is questioned. The possibility of weak “moral” AI scenarios is likewise discussed critically.

The Paradox of Patience

Some normative theories—act utilitarianism and rational decision theory among them—both designate some range of outcomes as particularly important, and, with an eye towards securing those outcomes, provide agents with advice concerning what is to be done. In this paper I argue that there are situations in which these two aspects of such theories are in tension. I provide a handful of conditions that, when jointly satisfi ed, pick out situations in which these theories recommend that agents act in ways that do not contribute to the outcomes on which they place value.

The Eligibility of Rule Utilitarianism

According to the eligibility theory of meaning, often attributed to David Lewis, the referent of a predicate is the property that best balances the twin constraints of charity (i.e. fit with our usage of the term) and eligibility, where eligibility is a function of metaphysical naturalness (i.e. how much of a natural kind the property is). This sort of metasemantics, which is motivated by its ability to resolve problems of indeterminacy and secure shared reference between disputing parties, can be somewhat friendly towards revisionary (i.e. counterintuitive) theories, since highly natural properties can act as “reference magnets,” securing our reference despite some mismatch with usage. In this paper, I apply these considerations to normative ethics and argue that the theory of rule utilitarianism achieves a high balance of charity and eligibility. I proceed by comparing rule utilitarianism to two of its well-known rivals, act utilitarianism and Rossian pluralism (a.k.a. “Commonsense Morality”). I show how the former achieves a high degree of eligibility but only at a significant cost of charity, while the latter does the opposite, fitting very nicely with our considered judgments but at the price of very low eligibility. Rule utilitarianism, on the other hand, strikes a good balance between these extremes; it assigns to our core moral term (‘moral permissibility’) a relatively natural property without doing too much damage to our moral convictions. Thus, rule utilitarianism should be regarded as a promising moral theory by any philosopher who takes seriously considerations of eligibility and naturalness.