I propose to consider the question, "Can machines think?" This should begin with definitions of the meaning of the terms "machine" and "think." The definitions might be framed so as to reflect so far as possible the normal use of the words, but this attitude is dangerous, If the meaning of the words "machine" and "think" are to be found by examining how they are commonly used it is difficult to escape the conclusion that the meaning and the answer to (...) the question, "Can machines think?" is to be sought in a statistical survey such as a Gallup poll. But this is absurd. Instead of attempting such a definition I shall replace the question by another, which is closely related to it and is expressed in relatively unambiguous words. The new form of the problem can be described in terms of a game which we call the 'imitation game." It is played with three people, a man (A), a woman (B), and an interrogator (C) who may be of either sex. The interrogator stays in a room apart front the other two. The object of the game for the interrogator is to determine which of the other two is the man and which is the woman. He knows them by labels X and Y, and at the end of the game he says either "X is A and Y is B" or "X is B and Y is A." The interrogator is allowed to put questions to A and B. We now ask the question, "What will happen when a machine takes the part of A in this game?" Will the interrogator decide wrongly as often when the game is played like this as he does when the game is played between a man and a woman? These questions replace our original, "Can machines think?". (shrink)

This article reports the findings of AI4People, an Atomium—EISMD initiative designed to lay the foundations for a “Good AI Society”. We introduce the core opportunities and risks of AI for society; present a synthesis of five ethical principles that should undergird its development and adoption; and offer 20 concrete recommendations—to assess, to develop, to incentivise, and to support good AI—which in some cases may be undertaken directly by national or supranational policy makers, while in others may be led by other (...) stakeholders. If adopted, these recommendations would serve as a firm foundation for the establishment of a Good AI Society. (shrink)

Value Sensitive Design (VSD) is an established method for integrating values into technical design. It has been applied to different technologies and, more recently, to artificial intelligence (AI). We argue that AI poses a number of challenges specific to VSD that require a somewhat modified VSD approach. Machine learning (ML), in particular, poses two challenges. First, humans may not understand how an AI system learns certain things. This requires paying attention to values such as transparency, explicability, and accountability. Second, ML (...) may lead to AI systems adapting in ways that ‘disembody’ the values embedded in them. To address this, we propose a threefold modified VSD approach: 1) integrating a known set of VSD principles (AI4SG) as design norms from which more specific design requirements can be derived; 2) distinguishing between values that are promoted and respected by the design to ensure outcomes that not only do no harm but also contribute to good; and 3) extending the VSD process to encompass the whole life cycle of an AI technology in order to monitor unintended value consequences and redesign as needed. We illustrate our VSD for AI approach with an example use case of a SARS-CoV-2 contact tracing app. (shrink)

In this much-anticipated book, Jonathan Dancy offers the only available full-scale treatment of particularism in ethics, a view with which he has been associated for twenty years. Dancy now presents particularism as the view that the possibility of moral thought and judgement does not in any way depend on an adequate supply of principles. He grounds this claim on a form of reasons-holism, holding that what is a reason in one case need not be any reason in another, and maintaining (...) that moral reasons are no different in this respect from others. Ethics Without Principles offers detailed treatment and criticism of opposing positions of all sorts, with useful summaries. It also contains three chapters on the nature of what Dancy calls 'contributory' reasons, which, though the topic of much interest at the moment, are not often discussed in anything like the detail given here. And it offers a distinctive form of value-holism to go with the holism of reasons. As Dancy's definitive statement on particularism, the book will be required reading for all those working on moral philosophy and ethical theory. (shrink)

A principal goal of the discipline of artificial morality is to design artificial agents to act as if they are moral agents. Intermediate goals of artificial morality are directed at building into AI systems sensitivity to the values, ethics, and legality of activities. The development of an effective foundation for the field of artificial morality involves exploring the technological and philosophical issues involved in making computers into explicit moral reasoners. The goal of this paper is to discuss strategies for implementing (...) artificial morality and the differing criteria for success that are appropriate to different strategies. (shrink)

The emerging discipline of Machine Ethics is concerned with creating autonomous artificial moral agents that perform ethically significant actions out in the world. Recently, Wallach and Allen (Moral machines: teaching robots right from wrong, Oxford University Press, Oxford, 2009) and others have argued that a virtue-based moral framework is a promising tool for meeting this end. However, even if we could program autonomous machines to follow a virtue-based moral framework, there are certain pressing ethical issues that need to be taken (...) into account, prior to the implementation and development stages. Here I examine whether the creation of virtuous autonomous machines is morally permitted by the central tenets of virtue ethics. It is argued that the creation of such machines violates certain tenets of virtue ethics, and hence that the creation and use of those machines is impermissible. One upshot of this is that, although virtue ethics may have a role to play in certain near-term Machine Ethics projects (e.g. designing systems that are sensitive to ethical considerations), machine ethicists need to look elsewhere for a moral framework to implement into their autonomous artificial moral agents, Wallach and Allen’s claims notwithstanding. (shrink)

This paper makes the non-monotonicity of a wide range of moral reasoning the basis of a case for particularism. Non-monotonicity threatens practical decision with an overwhelming informational complexity to which a form of ethical generalism seems the best response. It is argued that this impression is wholly misleading: the fact of non-monotonicity is best accommodated by the defence of four related theses in any theory of justification. First, the explanation of and defence of a default/challenge model of justification. Secondly, the (...) development of a theory of epistemic status and an explanation of those unearned entitlements that accrue to such status. Thirdly, an explanation of the basis of epistemic virtues. Finally, an account must be given of the executive capacity of rational decision itself as a ‘contentless ability’. This overall set of views can accommodate a limited role for generalizations about categories of evidence, but not such as to rescue a principled generalism. In particular, the version of particularism defended here explains why one ought not to accept the principled holism that has proved to be a problem for Dancy’s form of particularism. Ethics certainly involves hedged principles. However, principles cannot be self-hedging: there cannot be a that’s it operator in a principle as Richard Holton has claimed that there can be. Practical reasoning is concluded by the categorical detachment of the action-as-conclusion itself. (shrink)

How can we best identify, understand, and deal with ethical and societal issues raised by healthcare robotics? This paper argues that next to ethical analysis, classic technology assessment, and philosophical speculation we need forms of reflection, dialogue, and experiment that come, quite literally, much closer to innovation practices and contexts of use. The authors discuss a number of ways how to achieve that. Informed by their experience with “embedded” ethics in technical projects and with various tools and methods of responsible (...) research and innovation, the paper identifies “internal” and “external” forms of dialogical research and innovation, reflections on the possibilities and limitations of these forms of ethical–technological innovation, and explores a number of ways how they can be supported by policy at national and supranational level. (shrink)

There was a long tradition in philosophy according to which good reasoning had to be deductively valid. However, that tradition began to be questioned in the 1960’s, and is now thoroughly discredited. What caused its downfall was the recognition that many familiar kinds of reasoning are not deductively valid, but clearly confer justification on their conclusions. Here are some simple examples.

This paper employs (and defends where needed) a familiar four-part methodology for assessing moral theories. This methodology makes the most popular kind of moral pluralism--here called Ross-style pluralism--look extremely attractive. The paper contends, however, that, if rule-consequentialism's implications match our considered moral convictions as well as Ross-style pluralism's implications do, the methodology makes rule-consequentialism look even more attractive than Ross-style pluralism. The paper then attacks two arguments recently put forward in defence of Ross-style pluralism. One of these arguments is that (...) no moral theory containing some single normative principle to justify general pro tanto duties can do justice to the ineliminable role of judgment in moral thinking. The other argument is that no such theory is plausible in light of the fact that our moral ideas come from disparate historical sources. (shrink)

Should particularists about ethics claim that moral principles are never true? Or should they rather claim that any finite set of principles will not be sufficient to capture ethics? This paper explores and defends the possibility of embracing the second of these claims whilst rejecting the first, a position termed 'principled particularism'. The main argument that particularists present for their position-the argument that holds that any moral conclusion can be superseded by further considerations-is quite compatible with principled particularism; indeed, it (...) is compatible with the idea that every true moral conclusion can be shown to follow deductively from a finite set of premises. Whilst it is true that these premises must contain implicit ceteris paribus clauses, this does not render the arguments trivial. On the contrary, they can do important work in justifying moral conclusions. Finally the approach is briefly applied to the related field of jurisprudence. (shrink)

It seems that hope springs eternal for the cherished idea that norms (or normativeprinciples) explain actions or regularities in actions. But it also seems thatthere are many ways of going wrong when taking norms and normative principlesas explanatory. The author argues that neither norms nor normative principles—insofar as they are the sort of things with normative force—is explanatoryof what is done. He considers the matter using both erotetic and ontic models ofexplanation. He further considers various understandings of norms. Key Words: (...) explanation • norms • social science • rationality. (shrink)

Social robots are robots that can interact socially with humans. As social robots and the artificial intelligence that powers them becomes more advanced, they will likely take on more social and work roles. This has many important ethical implications. In this paper, we focus on one of the most central of these, the impacts that social robots can have on human autonomy. We argue that, due to their physical presence and social capacities, there is a strong potential for social robots (...) to enhance human autonomy as well as several ways they can inhibit and disrespect it. We argue that social robots could improve human autonomy by helping us to achieve more valuable ends, make more authentic choices, and improve our autonomy competencies. We also argue that social robots have the potential to harm human autonomy by instead leading us to achieve fewer valuable ends ourselves, make less authentic choices, decrease our autonomy competencies, make our autonomy more vulnerable, and disrespect our autonomy. Whether the impacts of social robots on human autonomy are positive or negative overall will depend on the design, regulation, and use we make of social robots in the future. (shrink)

Artificial agents (AAs), particularly but not only those in Cyberspace, extend the class of entities that can be involved in moral situations. For they can be conceived of as moral patients (as entities that can be acted upon for good or evil) and also as moral agents (as entities that can perform actions, again for good or evil). In this paper, we clarify the concept of agent and go on to separate the concerns of morality and responsibility of agents (most (...) interestingly for us, of AAs). We conclude that there is substantial and important scope, particularly in Computer Ethics, for the concept of moral agent not necessarily exhibiting free will, mental states or responsibility. This complements the more traditional approach, common at least since Montaigne and Descartes, which considers whether or not (artificial) agents have mental states, feelings, emotions and so on. By focussing directly on mind-less morality we are able to avoid that question and also many of the concerns of Artificial Intelligence. A vital component in our approach is the Method of Abstraction for analysing the level of abstraction (LoA) at which an agent is considered to act. The LoA is determined by the way in which one chooses to describe, analyse and discuss a system and its context. The Method of Abstraction is explained in terms of an interface or set of features or observables at a given LoA. Agenthood, and in particular moral agenthood, depends on a LoA. Our guidelines for agenthood are: interactivity (response to stimulus by change of state), autonomy (ability to change state without stimulus) and adaptability (ability to change the transition rules by which state is changed) at a given LoA. Morality may be thought of as a threshold defined on the observables in the interface determining the LoA under consideration. An agent is morally good if its actions all respect that threshold; and it is morally evil if some action violates it. That view is particularly informative when the agent constitutes a software or digital system, and the observables are numerical. Finally we review the consequences for Computer Ethics of our approach. In conclusion, this approach facilitates the discussion of the morality of agents not only in Cyberspace but also in the biosphere, where animals can be considered moral agents without their having to display free will, emotions or mental states, and in social contexts, where systems like organizations can play the role of moral agents. The primary cost of this facility is the extension of the class of agents and moral agents to embrace AAs. (shrink)

This paper offers a general model of substantive moral principles as a kind of hedged moral principles that can (but don't have to) tolerate exceptions. I argue that the kind of principles I defend provide an account of what would make an exception to them permissible. I also argue that these principles are nonetheless robustly explanatory with respect to a variety of moral facts; that they make sense of error, uncertainty, and disagreement concerning moral principles and their implications; and that (...) one can grasp these principles without having to grasp any particular list of their permissibly exceptional instances. I conclude by pointing out various advantages that this model of principles has over several of its rivals. The bottom line is that we should find nothing peculiarly odd or problematic about the idea of exception-tolerating and yet robustly explanatory moral principles. (shrink)

There are at least three things we might mean by "ethics in robotics": the ethical systems built into robots, the ethics of people who design and use robots, and the ethics of how people treat robots. This paper argues that the best approach to robot ethics is one which addresses all three of these, and to do this it ought to consider robots as socio-technical systems. By so doing, it is possible to think of a continuum of agency that lies (...) between amoral and fully autonomous moral agents. Thus, robots might move gradually along this continuum as they acquire greater capabilities and ethical sophistication. It also argues that many of the issues regarding the distribution of responsibility in complex socio-technical systems might best be addressed by looking to legal theory, rather than moral theory. This is because our overarching interest in robot ethics ought to be the practical one of preventing robots from doing harm, as well as preventing humans from unjustly avoiding responsibility for their actions. (shrink)

This paper argues that although moral intuitions are insufficient for making judgments on new technological innovations, they maintain great utility for informing responsible innovation. To do this, this paper employs the Value Sensitive Design (VSD) methodology as an illustrative example of how stakeholder values can be better distilled to inform responsible innovation. Further, it is argued that moral intuitions are necessary for determining stakeholder values required for the design of responsible technologies. This argument is supported by the claim that the (...) moral intuitions of stakeholders allow designers to conceptualize stakeholder values and incorporate them into the early phases of design. It is concluded that design-for-values (DFV) frameworks like the VSD methodology can remain potent if developers adopt heuristic tools to diminish the influence of cognitive biases thus strengthening the reliability of moral intuitions. (shrink)

This paper argues that the Value Sensitive Design (VSD) methodology provides a principled approach to embedding common values in to AI systems both early and throughout the design process. To do so, it draws on an important case study: the evidence and final report of the UK Select Committee on Artificial Intelligence. This empirical investigation shows that the different and often disparate stakeholder groups that are implicated in AI design and use share some common values that can be used to (...) further strengthen design coordination efforts. VSD is shown to be both able to distill these common values as well as provide a framework for stakeholder coordination. (shrink)

The goal of this paper is to frame a theory of reasons--what they are, how they support actions or conclusions--using the tools of default logic. After sketching the basic account of reasons as provided by defaults, I show how it can be elaborated to deal with two more complicated issues: first, situations in which the priority relation among defaults, and so reasons as well, is itself established through default reasoning; second, the treatment of undercutting defeat and exclusionary reasons. Finally, and (...) by way of application, I show how the resulting account can shed some light on Jonathan Dancy's argument from reason holism to a form of extreme particularism in moral theory. (shrink)