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)

Traditionally, the manufacturer/operator of a machine is held (morally and legally) responsible for the consequences of its operation. Autonomous, learning machines, based on neural networks, genetic algorithms and agent architectures, create a new situation, where the manufacturer/operator of the machine is in principle not capable of predicting the future machine behaviour any more, and thus cannot be held morally responsible or liable for it. The society must decide between not using this kind of machine any more (which is not a (...) realistic option), or facing a responsibility gap, which cannot be bridged by traditional concepts of responsibility ascription. (shrink)

The intention of this paper is to point to the dilemma humanity may face in light of AI advancements. The dilemma is whether to create a world with less evil or maintain the human status of moral agents. This dilemma may arise as a consequence of using automated decision-making systems for high-stakes decisions. The use of automated decision-making bears the risk of eliminating human moral agency and autonomy and reducing humans to mere moral patients. On the other hand, it also (...) has the potential to bring tremendous benefits to humanity by decreasing human-induced harm in the world. After presenting how this dilemma may arise, I explore general avenues for addressing it. I will argue that we do not have to solve this dilemma in an all-or-nothing fashion and that a more nuanced approach may be suitable. However, the main point I want to highlight is that we need to have a principled way of addressing this dilemma, which is currently missing. (shrink)

There is a concern that the widespread deployment of autonomous machines will open up a number of ‘responsibility gaps’ throughout society. Various articulations of such techno-responsibility gaps have been proposed over the years, along with several potential solutions. Most of these solutions focus on ‘plugging’ or ‘dissolving’ the gaps. This paper offers an alternative perspective. It argues that techno-responsibility gaps are, sometimes, to be welcomed and that one of the advantages of autonomous machines is that they enable us to embrace (...) certain kinds of responsibility gap. The argument is based on the idea that human morality is often tragic. We frequently confront situations in which competing moral considerations pull in different directions and it is impossible to perfectly balance these considerations. This heightens the burden of responsibility associated with our choices. We cope with the tragedy of moral choice in different ways. Sometimes we delude ourselves into thinking the choices we make were not tragic ; sometimes we delegate the tragic choice to others ; sometimes we make the choice ourselves and bear the psychological consequences. Each of these strategies has its benefits and costs. One potential advantage of autonomous machines is that they enable a reduced cost form of delegation. However, we only gain the advantage of this reduced cost if we accept that some techno-responsibility gaps are virtuous. (shrink)

Calls to hold artificial intelligence to account are intensifying. Activists and researchers alike warn of an “accountability gap” or even a “crisis of accountability” in AI. Meanwhile, several prominent scholars maintain that accountability holds the key to governing AI. But usage of the term varies widely in discussions of AI ethics and governance. This chapter begins by disambiguating some different senses and dimensions of accountability, distinguishing it from neighboring concepts, and identifying sources of confusion. It proceeds to explore the idea (...) that AI operates within an accountability gap arising from technical features of AI as well as the social context in which it is deployed. The chapter also evaluates various proposals for closing this gap. I conclude that the role of accountability in AI ethics and governance is vital but also more limited than some suggest. Accountability’s primary job description is to verify compliance with substantive normative principles—once those principles are settled. Theories of accountability cannot ultimately tell us what substantive standards to account for, especially when norms are contested or still emerging. Nonetheless, formal mechanisms of accountability provide a way of diagnosing and discouraging egregious wrongdoing even in the absence of normative agreement. Providing accounts can also be an important first step toward the development of more comprehensive regulatory standards for AI. (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)

Certain characteristics make machine learning a powerful tool for processing large amounts of data, and also particularly unsuitable for explanatory purposes. There are worries that its increasing use in science may sideline the explanatory goals of research. We analyze the key characteristics of ML that might have implications for the future directions in scientific research: epistemic opacity and the ‘theory-agnostic’ modeling. These characteristics are further analyzed in a comparison of ML with the traditional statistical methods, in order to demonstrate what (...) it is specifically that makes ML methodology substantially unsuitable for reaching explanations. The analysis is given broader philosophical context by connecting it with the views on the role of prediction and explanation in science, their relationship, and the value of explanation. We proceed to show, first, that ML disrupts the relationship between prediction and explanation commonly understood as a functional relationship. Then we show that the value of explanation is not exhausted in purely predictive functions, but rather has a ubiquitously recognized value for both science and everyday life. We then invoke two hypothetical scenarios with different degrees of automatization of science, which help test our intuitions on the role of explanation in science. The main question we address is whether ML will reorient or otherwise impact our standard explanatory practice. We conclude with a prognosis that ML would diversify science into purely predictively oriented research based on ML-like techniques and, on the other hand, remaining faithful to anthropocentric research focused on the search for explanation. (shrink)

Are groups ever capable of bearing responsibility, over and above their individual members? This chapter discusses and defends the view that certain organized collectives – namely, those that qualify as group moral agents – can be held responsible for their actions, and that group responsibility is not reducible to individual responsibility. The view has important implications. It supports the recognition of corporate civil and even criminal liability in our legal systems, and it suggests that, by recognizing group agents as loci (...) of responsibility, we may be able to avoid “responsibility gaps” in some cases of collectively caused harms for which there is a shortfall of individual responsibility. The chapter further asks whether the view that certain groups are responsible agents commits us to the view that those groups should also be given rights of their own and gives a qualified negative answer. (shrink)

The notion of “responsibility gap” with artificial intelligence (AI) was originally introduced in the philosophical debate to indicate the concern that “learning automata” may make more difficult or impossible to attribute moral culpability to persons for untoward events. Building on literature in moral and legal philosophy, and ethics of technology, the paper proposes a broader and more comprehensive analysis of the responsibility gap. The responsibility gap, it is argued, is not one problem but a set of at least four interconnected (...) problems – gaps in culpability, moral and public accountability, active responsibility—caused by different sources, some technical, other organisational, legal, ethical, and societal. Responsibility gaps may also happen with non-learning systems. The paper clarifies which aspect of AI may cause which gap in which form of responsibility, and why each of these gaps matter. It proposes a critical review of partial and non-satisfactory attempts to address the responsibility gap: those which present it as a new and intractable problem (“fatalism”), those which dismiss it as a false problem (“deflationism”), and those which reduce it to only one of its dimensions or sources and/or present it as a problem that can be solved by simply introducing new technical and/or legal tools (“solutionism”). The paper also outlines a more comprehensive approach to address the responsibility gaps with AI in their entirety, based on the idea of designing socio-technical systems for “meaningful human control", that is systems aligned with the relevant human reasons and capacities. (shrink)

It has been argued that we cannot be morally responsible in the sense required to deserve blame or punishment if the world is deterministic, but still morally responsible in the sense of being apt targets for moral criticism. Desert-entailing moral responsibility is supposed to be more freedom-demanding than other kinds of responsibility, since it justifies subjecting people to blame and punishments, is nonconsequentialist, and has been shown by thought experiments to be incompatible with determinism. In this paper, I will show (...) that all these arguments can be resisted. (shrink)

In this important collection of essays Dennis Thompson argues for a more robust conception of responsibility in public life than prevails in contemporary democracies. He suggests that we should stop thinking so much about public ethics in terms of individual vices and start thinking about it more in terms of institutional vices. Combining theory and practice with many concrete examples and proposals for reform, these essays could be used in courses in applied ethics or political theory and will be read (...) by professionals and graduate students in schools of political science, public policy, law, public health, journalism and business. (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)

Utilitarianism, the great reforming philosophy of the nineteenth century, has today acquired the reputation for being a crassly calculating, impersonal philosophy unfit to serve as a guide to moral conduct. Yet what may disqualify utilitarianism as a personal philosophy makes it an eminently suitable guide for public officials in the pursuit of their professional responsibilities. Robert E. Goodin, a philosopher with many books on political theory, public policy and applied ethics to his credit, defends utilitarianism against its critics and shows (...) how it can be applied most effectively over a wide range of public policies. In discussions of such issues as paternalism, social welfare policy, international ethics, nuclear armaments, and international responses to the environment crisis, he demonstrates what a flexible tool his brand of utilitarianism can be in confronting the dilemmas of public policy in the real world. (shrink)

David Shoemaker presents a new pluralistic theory of responsibility, based on the idea of quality of will. His approach is motivated by our ambivalence to real-life cases of marginal agency, such as those caused by clinical depression, dementia, scrupulosity, psychopathy, autism, intellectual disability, and poor formative circumstances. Our ambivalent responses suggest that such agents are responsible in some ways but not others. Shoemaker develops a theory to account for our ambivalence, via close examination of several categories of pancultural emotional responsibility (...) responses and their appropriateness conditions. The result is three distinct types of responsibility, each with its own set of required capacities: attributability, answerability, and accountability. Attributability is about the having and expressing of various traits of character, and it is the target of a range of aretaic sentiments and emotional practices organized around disdain and admiration. Answerability is about one’s capacity to govern one’s actions and attitudes by one’s evaluative judgments about the worth of various practical reasons, and it is the target of a range of sentiments and emotional practices organized around regret and pride. Accountability is about one’s ability to regard others, both evaluatively and emotionally, and it is the target of a range of sentiments and emotional practices organized around anger and gratitude. In Part One of the book, this tripartite theory is developed and defended. In Part Two of the book, the tripartite theory’s predictions about specific marginal cases are tested, once certain empirical details about the nature of those agents have been filled in and discussed. (shrink)

Artificially intelligent systems (AISs) are being created by software developing companies (SDCs) to influence clinical decision‐making. Historically, clinicians have led healthcare decision‐making, and the introduction of AISs makes SDCs novel actors in the clinical decision‐making space. Although these AISs are intended to influence a clinician's decision‐making, SDCs have been clear that clinicians are in fact the final decision‐makers in clinical care, and that AISs can only inform their decisions. As such, the default position is that clinicians should hold responsibility for (...) the outcomes of the use of AISs. This is not the case when an AIS has influenced a clinician's judgement and their subsequent decision. In this paper, we argue that this is an imbalanced and unjust position, and that careful thought needs to go into how personal moral responsibility for the use of AISs in clinical decision‐making should be attributed. This paper employs and examines the difference between prospective and retrospective responsibility and considers foreseeability as key in determining how personal moral responsibility can be justly attributed. This leads us to the view that moral responsibility for the outcomes of using AISs in healthcare ought to be shared by the clinical users and SDCs. (shrink)

This article considers the issue of opacity as a problem for socially consequential mechanisms of classification and ranking, such as spam filters, credit card fraud detection, search engines, news trends, market segmentation and advertising, insurance or loan qualification, and credit scoring. These mechanisms of classification all frequently rely on computational algorithms, and in many cases on machine learning algorithms to do this work. In this article, I draw a distinction between three forms of opacity: opacity as intentional corporate or state (...) secrecy, opacity as technical illiteracy, and an opacity that arises from the characteristics of machine learning algorithms and the scale required to apply them usefully. The analysis in this article gets inside the algorithms themselves. I cite existing literatures in computer science, known industry practices, and do some testing and manipulation of code as a form of lightweight code audit. I argue that recognizing the distinct forms of opacity that may be coming into play in a given application is a key to determining which of a variety of technical and non-technical solutions could help to prevent harm. (shrink)

Philosophers usually discuss responsibility in terms of responsibility for past actions or as a question about the nature of moral agency. Yet the word responsibility is fairly modern, whereas these topics arguably represent timeless concerns about human agency. This paper investigates another use of responsibility, that is particularly important to modern liberal societies: responsibility as a virtue that can be demonstrated by individuals and organisations. The paper notes its initial importance in political contexts, and seeks to explain why we now (...) demand responsibility in all spheres of life. In reply, I highlight the distinctively institutional character of modern liberal societies: institutions specify many of the particular responsibilities each of us must fulfil, but also require responsibility to sustain them and address their failings. My overall argument is that the virtue of responsibility occupies a distinctive place in the moral needs, and moral achievements, of liberal societies; and this, in turn, explains why it now occupies such a prominent place in our moral discourse. (shrink)

Because of its practical advantages, machine learning (ML) is increasingly used for decision-making in numerous sectors. This paper demonstrates that the integral characteristics of ML, such as semi-autonomy, complexity, and non-deterministic modeling have important ethical implications. In particular, these characteristics lead to a lack of insight and lack of comprehensibility, and ultimately to the loss of human control over decision-making. Errors, which are bound to occur in any decision-making process, may lead to great harm and human rights violations. It is (...) important to have a principled way of assigning responsibility for such errors. The integral characteristics of ML, however, pose serious difficulties in defining responsibility and regulating ML decision-making. First, we elaborate on these characteristics and their epistemic and ethical implications. We then analyze possible general strategies for resolving the assignment of moral responsibility and show that, due to the specific way in which ML functions, each potential solution is problematic, whether we assign responsibility to humans, machines, or using hybrid models. Then, we shift focus on an alternative approach that bypasses moral responsibility and attempts to define legal liability independently through solutions such as informed consent and the no-fault compensation system. Both of these solutions prove unsatisfactory because they leave too much room for potential abuses of ML decision-making. We conclude that both ethical and legal solutions are fraught with serious difficulties. These difficulties prompt us to re-weigh the costs and benefits of using ML for high-stake decisions. (shrink)

The notion of responsible innovation suggests that innovators carry additional responsibilities beyond those commonly suggested. In this paper, we will discuss the meaning of these novel responsibilities focusing on two philosophical problems of attributing such responsibilities to innovators. The first is the allocation of responsibilities to innovators. Innovation is a process that involves a multiplicity of agents and unpredictable, far-reaching causal chains from innovation to social impacts, which creates great uncertainty. A second problem is constituted by possible trade-offs between different (...) kinds of responsibility. It is evident that attributing backward-looking responsibility for product failures diminishes the willingness to learn about such defects and to take forward-looking responsibility. We will argue that these problems can be overcome by elaborating what it is exactly that innovators are responsible for. In this manner, we will distinguish more clearly between holding responsible and taking responsibility. This opens a space for ‘supererogatory’ responsibilities. Second, we will argue that both innovation processes and outcomes can be objects of innovators’ responsibility. Third, we will analyze different kinds of responsibility and show that the functions of their attribution are not necessarily contradictory. Based on this conceptual refinement, we will argue that accountability, responsibility-as-virtue and the willingness to take responsibility are crucial for responsible innovation. (shrink)

The aim of this literature review was to compose a narrative review supported by a systematic approach to critically identify and examine concerns about accountability and the allocation of responsibility and legal liability as applied to the clinician and the technologist as applied the use of opaque AI-powered systems in clinical decision making. This review questions if it is permissible for a clinician to use an opaque AI system in clinical decision making and if a patient was harmed as a (...) result of using a clinician using an AIS’s suggestion, how would responsibility and legal liability be allocated? Literature was systematically searched, retrieved, and reviewed from nine databases, which also included items from three clinical professional regulators, as well as relevant grey literature from governmental and non-governmental organisations. This literature was subjected to inclusion/exclusion criteria; those items found relevant to this review underwent data extraction. This review found that there are multiple concerns about opacity, accountability, responsibility and liability when considering the stakeholders of technologists and clinicians in the creation and use of AIS in clinical decision making. Accountability is challenged when the AIS used is opaque, and allocation of responsibility is somewhat unclear. Legal analysis would help stakeholders to understand their obligations and prepare should an undesirable scenario of patient harm eventuate when AIS were used. (shrink)

Automated technologies and robots make decisions that cannot always be fully controlled or predicted. In addition to that, they cannot respond to punishment and blame in the ways humans do. Therefore, when automated cars harm or kill people, for example, this gives rise to concerns about responsibility-gaps and retribution-gaps. According to Sven Nyholm, however, automated cars do not pose a challenge on human responsibility, as long as humans can control them and update them. He argues that the agency exercised in (...) automated cars should be understood in terms of human–robot collaborations. This brief note focuses on the problem that arises when there are multiple people involved, but there is no obvious shared collaboration among them. Building on John Danaher’s discussion of command responsibility, it is argued that, although Nyholm might be right that autonomous cars cannot be regarded as acting on their own, independently of any human beings, worries about responsibility-gaps and retribution-gaps are still justified, because it often remains unclear how to allocate or distribute responsibility satisfactorily among the key humans involved after they have been successfully identified. (shrink)