Predictive algorithms are playing an increasingly prominent role in society, being used to predict recidivism, loan repayment, job performance, and so on. With this increasing influence has come an increasing concern with the ways in which they might be unfair or biased against individuals in virtue of their race, gender, or, more generally, their group membership. Many purported criteria of algorithmic fairness concern statistical relationships between the algorithm’s predictions and the actual outcomes, for instance requiring that the rate of false (...) positives be equal across the relevant groups. We might seek to ensure that algorithms satisfy all of these purported fairness criteria. But a series of impossibility results shows that this is impossible, unless base rates are equal across the relevant groups. What are we to make of these pessimistic results? I argue that none of the purported criteria, except for a calibration criterion, are necessary conditions for fairness, on the grounds that they can all be simultaneously violated by a manifestly fair and uniquely optimal predictive algorithm, even when base rates are equal. I conclude with some general reflections on algorithmic fairness. (shrink)

We are living through an era of increased robotisation. Some authors have already begun to explore the impact of this robotisation on legal rules and practice. In doing so, many highlight potential liability gaps that might arise through robot misbehaviour. Although these gaps are interesting and socially significant, they do not exhaust the possible gaps that might be created by increased robotisation. In this article, I make the case for one of those alternative gaps: the retribution gap. This gap arises (...) from a mismatch between the human desire for retribution and the absence of appropriate subjects of retributive blame. I argue for the potential existence of this gap in an era of increased robotisation; suggest that it is much harder to plug this gap than it is to plug those thus far explored in the literature; and then highlight three important social implications of this gap. (shrink)

The United States Army’s Future Combat Systems Project, which aims to manufacture a “robot army” to be ready for deployment by 2012, is only the latest and most dramatic example of military interest in the use of artificially intelligent systems in modern warfare. This paper considers the ethics of a decision to send artificially intelligent robots into war, by asking who we should hold responsible when an autonomous weapon system is involved in an atrocity of the sort that would normally (...) be described as a war crime. A number of possible loci of responsibility for robot war crimes are canvassed; the persons who designed or programmed the system, the commanding officer who ordered its use, the machine itself. I argue that in fact none of these are ultimately satisfactory. Yet it is a necessary condition for fighting a just war, under the principle of jus in bellum, that someone can be justly held responsible for deaths that occur in the course of the war. As this condition cannot be met in relation to deaths caused by an autonomous weapon system it would therefore be unethical to deploy such systems in warfare. (shrink)

Recent decades have witnessed tremendous progress in artificial intelligence and in the development of autonomous systems that rely on artificial intelligence. Critics, however, have pointed to the difficulty of allocating responsibility for the actions of an autonomous system, especially when the autonomous system causes harm or damage. The highly autonomous behavior of such systems, for which neither the programmer, the manufacturer, nor the operator seems to be responsible, has been suspected to generate responsibility gaps. This has been the cause of (...) much concern. In this article, I propose a more optimistic view on artificial intelligence, raising two challenges for responsibility gap pessimists. First, proponents of responsibility gaps must say more about when responsibility gaps occur. Once we accept a difficult-to-reject plausibility constraint on the emergence of such gaps, it becomes apparent that the situations in which responsibility gaps occur are unclear. Second, assuming that responsibility gaps occur, more must be said about why we should be concerned about such gaps in the first place. I proceed by defusing what I take to be the two most important concerns about responsibility gaps, one relating to the consequences of responsibility gaps and the other relating to violations of jus in bello. (shrink)

Previous research in Explainable Artificial Intelligence (XAI) suggests that a main aim of explainability approaches is to satisfy specific interests, goals, expectations, needs, and demands regarding artificial systems (we call these “stakeholders' desiderata”) in a variety of contexts. However, the literature on XAI is vast, spreads out across multiple largely disconnected disciplines, and it often remains unclear how explainability approaches are supposed to achieve the goal of satisfying stakeholders' desiderata. This paper discusses the main classes of stakeholders calling for explainability (...) of artificial systems and reviews their desiderata. We provide a model that explicitly spells out the main concepts and relations necessary to consider and investigate when evaluating, adjusting, choosing, and developing explainability approaches that aim to satisfy stakeholders' desiderata. This model can serve researchers from the variety of different disciplines involved in XAI as a common ground. It emphasizes where there is interdisciplinary potential in the evaluation and the development of explainability approaches. (shrink)

We argue that explainable artificial intelligence (XAI), specifically reason-giving XAI, often constitutes the most suitable way of ensuring that someone can properly be held responsible for decisions that are based on the outputs of artificial intelligent (AI) systems. We first show that, to close moral responsibility gaps (Matthias 2004), often a human in the loop is needed who is directly responsible for particular AI-supported decisions. Second, we appeal to the epistemic condition on moral responsibility to argue that, in order to (...) be responsible for her decision, the human in the loop has to have an explanation available of the system’s recommendation. Reason explanations are especially well-suited to this end and we examine whether – and how – it might be possible to make such explanations fit with AI systems. We support our claims by focusing on a case of disagreement between human in the loop and AI system. (shrink)

Philosophy & Public Affairs, Volume 50, Issue 2, Page 239-266, Spring 2022.

An encyclopedia article on the epistemic or knowledge condition for moral responsibility, written for the SEP.

There is increasing speculation within military and policy circles that the future of armed conflict is likely to include extensive deployment of robots designed to identify targets and destroy them without the direct oversight of a human operator. My aim in this paper is twofold. First, I will argue that the ethical case for allowing autonomous targeting, at least in specific restricted domains, is stronger than critics have acknowledged. Second, I will attempt to uncover, explicate, and defend the intuition that (...) even in this context there would be something ethically problematic about such targeting. I argue that an account of the non-consequentialist foundations of the principle of distinction suggests that the use of autonomous weapon systems is unethical by virtue of failing to show appropriate respect for the humanity of our enemies. However, the success of the strongest form of this argument depends upon understanding the robot itself as doing the killing. To the extent that we believe that, on the contrary, AWS are only the means whereby those who order them into action kill, the idea that the use of AWS fails to respect the humanity of our enemy will turn upon an account of what is required by respect that is essentially conventional. Thus, while the theoretical foundations of the idea that AWS are weapons that are “evil in themselves” are weaker than critics have sometimes maintained, they are nonetheless sufficient to the task of demanding a prohibition of the development and deployment of such weapons. (shrink)

The moral importance of liability to harm has so far been ignored in the lively debate about what self-driving vehicles should be programmed to do when an accident is inevitable. But liability matters a great deal to just distribution of risk of harm. While morality sometimes requires simply minimizing relevant harms, this is not so when one party is liable to harm in virtue of voluntarily engaging in activity that foreseeably creates a risky situation, while having reasonable alternatives. On plausible (...) assumptions, merely choosing to use a self-driving vehicle typically gives rise to a degree of liability, so that such vehicles should be programmed to shift the risk from bystanders to users, other things being equal. Insofar vehicles cannot be programmed to take all the factors affecting liability into account, there is a pro tanto moral reason not to introduce them, or restrict their use. (shrink)

This article offers an overview of the main first-order ethical questions raised by robots and Artificial Intelligence (RAIs) under five broad rubrics: functionality, inherent significance, rights and responsibilities, side-effects, and threats. The first letter of each rubric taken together conveniently generates the acronym FIRST. Special attention is given to the rubrics of functionality and inherent significance given the centrality of the former and the tendency to neglect the latter in virtue of its somewhat nebulous and contested character. In addition to (...) exploring some illustrative issues arising under each rubric, the article also emphasizes a number of more general themes. These include: the multiplicity of interacting levels on which ethical questions about RAIs arise, the need to recognise that RAIs potentially implicate the full gamut of human values (rather than exclusively or primarily some readily identifiable sub-set of ethical or legal principles), and the need for practically salient ethical reflection on RAIs to be informed by a realistic appreciation of their existing and foreseeable capacities. -/- . (shrink)

There has recently been increasing interest in the possibility and ethics of lethal autonomous weapons systems (LAWS), which would combine sophisticated AI with machinery capable of deadly force. One objection to LAWS is that their use will create a troubling responsibility gap, where no human agent can properly be held accountable for the outcomes that they create. While some authors have attempted to show that individual agents can, in fact, be responsible for the behaviour of LAWS in various circumstances, this (...) will not be possible in all cases. This article argues, however, that by assigning responsibilities to group agents operating within the military‐industrial complex, as well as the military‐industrial complex as a whole, we may be able to close the responsibility gap, or at least reduce its costs. Insofar as the potential for responsibility gaps poses an ethical barrier to the development and deployment of LAWS, this argument, if sound, may show their use to be morally permissible in a greater range of cases. (shrink)

Advances in artificial intelligence research allow us to build fairly sophisticated agents: robots and computer programs capable of acting and deciding on their own. These systems raise questions about who is responsible when something goes wrong—when such systems harm or kill humans. In a recent paper, Sven Nyholm has suggested that, because current AI will likely possess what we might call “supervised agency”, the theory of responsibility for individual agency is the wrong place to look for an answer to the (...) question of responsibility. Instead, or so argues Nyholm, because supervised agency is a form of collaborative agency—of acting together—the right place to look is the theory of collaborative responsibility—responsibility in cases of acting together. This paper concedes that current AI will possess supervised agency, but argues that it is nevertheless wrong to think of the relevant human-AI interactions as a form of collaborative agency and, hence, that responsibility in cases of collaborative agency is not the right place to look for the responsibility-grounding relation in human-AI interactions. It also suggests that the right place to look for this responsibility-grounding relation in human-AI interactions is the use of certain sorts of agents as instruments. (shrink)

Plans to automate killing by using robots armed with lethal weapons have been a prominent feature of most US military forces? roadmaps since 2004. The idea is to have a staged move from ?man-in-the-loop? to ?man-on-the-loop? to full autonomy. While this may result in considerable military advantages, the policy raises ethical concerns with regard to potential breaches of International Humanitarian Law, including the Principle of Distinction and the Principle of Proportionality. Current applications of remote piloted robot planes or drones offer (...) lessons about how automated weapons platforms could be misused by extending the range of legally questionable, targeted killings by security and intelligence forces. Moreover, the alleged moral disengagement by remote pilots will only be exacerbated by the use of autonomous robots. Leaders in the international community need to address the difficult legal and moral issues now, before the current mass proliferation of development reaches fruition. (shrink)

The introduction of Autonomous Military Systems (AMS) onto contemporary battlefields raises concerns that they will bring with them the possibility of a techno-responsibility gap, leaving insecurity about how to attribute responsibility in scenarios involving these systems. In this work I approach this problem in the domain of applied ethics with foundational conceptual work on autonomy and responsibility. I argue that concerns over the use of AMS can be assuaged by recognising the richly interrelated context in which these systems will most (...) likely be deployed. This will allow us to move beyond the solely individualist understandings of responsibility at work in most treatments of these cases, toward one that includes collective responsibility. This allows us to attribute collective responsibility to the collectives of which the AMS form a part, and to account for the distribution of burdens that follows from this attribution. I argue that this expansion of our responsibility practices will close at least some otherwise intractable techno-responsibility gaps. (shrink)

The Laws of Armed Conflict require that war crimes be attributed to individuals who can be held responsible and be punished. Yet assigning responsibility for the actions of Lethal Autonomous Weapon...

Serena Olsaretti brings together new essays by leading moral and political philosophers on the nature of desert and justice, their relations with each other and with other values.

This paper situates the problem of defeaters in a larger debate about the source of normative authority. It argues in favour of a constructivist account of defeasibility, which appeals to the justificatory role of moral principles. The argument builds upon the critique of two recent attempts to deal with defeasibility: first, a particularist account, which disposes of moral principles on the ground that reasons are holistic; and second, a proceduralist view, which addresses the problem of defeaters by distinguishing between provisional (...) and strictly universal principles. The particularist view fails to establish that moral principles have no epistemological import, but it raises important questions about their role in practical reasoning. The proceduralist view fails to distinguish between reasoning by default and reasoning by principles, but it shows that moral principles have a structural justificatory role. The constructivist view recognizes that the moral valence of normative claims vary across contexts, but denies that this is because of holism about reasons. Rather, it defends defeasibility within a constructivist account of reasoning where universality serves as the matrix of judgment. The constructivist view vindicates the justificatory role of universal moral principles, and makes room for some ordinary sources of defeasibility. (shrink)