Machine Ethics

A classification of the global catastrophic risks of AI is presented, along with a comprehensive list of previously identified risks. This classification allows the identification of several new risks. We show that at each level of AI’s intelligence power, separate types of possible catastrophes dominate. Our classification demonstrates that the field of AI risks is diverse, and includes many scenarios beyond the commonly discussed cases of a paperclip maximizer or robot-caused unemployment. Global catastrophic failure could happen at various levels of (...) AI development, namely, before it starts self-improvement, during its takeoff, when it uses various instruments to escape its initial confinement, or after it successfully takes over the world and starts to implement its goal system, which could be plainly unaligned, or feature-flawed friendliness. AI could also halt at later stages of its development either due to technical glitches or ontological problems. Overall, we identified around several dozen scenarios of AI-driven global catastrophe. The extent of this list illustrates that there is no one simple solution to the problem of AI safety, and that AI safety theory is complex and must be customized for each AI development level. (shrink)

An ambitious ethical theory ---Alan Gewirth's "Principle of Generic Consistency"--- is encoded and analysed in Isabelle/HOL. Gewirth's theory has stirred much attention in philosophy and ethics and has been proposed as a potential means to bound the impact of artificial general intelligence.

In this essay I try to determine the extent to which it is possible to conceive robots and organisms as analogous entities. After a cursory preamble on the long history of epistemological connections between machines and organisms I focus on Norbert Wiener’s cybernetics, where the analogy between modern machines and organisms is introduced most explicitly. The analysis of issues pertaining to the cybernetic interpretation of the analogy serves then as a basis for a critical assessment of its reprise in contemporary (...) robotics and AI, where the line between organisms and technologies gets even more blurred. In brief, I argue for the necessity of defining the extent to which the analogy between robots and organisms applies in order not only to harness its heuristic potential, but also to keep track of the pivotal differences that distinguish the two classes of entities. Finally, I discuss how the ordinary use of language and the peculiar status of the idea of robot obstruct awareness of the differences between robots and organisms. (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)

Given that artificial moral agents—such as autonomous vehicles, lethal autonomous weapons, and automated financial trading systems—are now part of the socio-ethical equation, we should morally evaluate their behavior. How should artificial moral agents make decisions? Is one moral theory better suited than others for machine ethics? After briefly overviewing the dominant ethical approaches for building morality into machines, this paper discusses a recent proposal, put forward by Don Howard and Ioan Muntean (2016, 2017), for an artificial moral agent based on (...) virtue theory. While the virtuous artificial moral agent has various strengths, this paper argues that a rule-based utilitarian approach (in contrast to a strict act-utilitarian approach) is superior because it can capture the most important features of the virtue-theoretic approach while realizing additional significant benefits. Specifically, a 2-level utilitarian artificial moral agent incorporating both established moral rules and a utility calculator is especially well-suited for machine ethics. (shrink)

The thinkers who have reflected on the problem of a coming superintelligence have generally seen the issue as a technological problem, a problem of how to control what the superintelligence will do. I argue that this approach is probably mistaken because it is based on questionable assumptions about the behavior of intelligent agents and, moreover, potentially counterproductive because it might, in the end, bring about the existential catastrophe that it is meant to prevent. I contend that the problem posed by (...) a future superintelligence will likely be a political problem, that is, one of establishing a peaceful form of coexistence with other intelligent agents in a situation of mutual vulnerability, and not a technological problem of control. (shrink)

Engineering an artificial intelligence to play an advisory role in morally charged decision making will inevitably introduce meta-ethical positions into the design. Some of these positions, by informing the design and operation of the AI, will introduce risks. This paper offers an analysis of these potential risks along the realism/anti-realism dimension in metaethics and reveals that realism poses greater risks, but, on the other hand, anti-realism undermines the motivation for engineering a moral AI in the first place.

The extent to which machine metaphors are used in synthetic biology is striking. These metaphors contain a specific perspective on organisms as well as on scientific and technological progress. Expressions such as “genetically engineered machine”, “genetic circuit”, and “platform organism”, taken from the realms of electronic engineering, car manufacturing, and information technology, highlight specific aspects of the functioning of living beings while at the same time hiding others, such as evolutionary change and interdependencies in ecosystems. Since these latter aspects are (...) relevant for, for example, risk evaluation of uncontained uses of synthetic organisms, it is ethically imperative to resist the thrust of machine metaphors in this respect. In addition, from the perspective of the machine metaphor viewing an entity as a moral agent or patient becomes dubious. If one were to regard living beings, including humans, as machines, it becomes difficult to justify ascriptions of moral status. Finally, the machine metaphor reinforces beliefs in the potential of synthetic biology to play a decisive role in solving societal problems, and downplays the role of alternative technological, and social and political measures. (shrink)

Artificial Life has two goals. One attempts to describe fundamental qualities of living systems through agent based computer models. And the second studies whether or not we can artificially create living things in computational mediums that can be realized either, virtually in software, or through biotechnology. The study of ALife has recently branched into two further subdivisions, one is “dry” ALife, which is the study of living systems “in silico” through the use of computer simulations, and the other is “wet” (...) ALife that uses biological material to realize what has only been simulated on computers, effectively wet ALife uses biological material as a kind of computer. This is challenging to the field of computer ethics as it points towards a future in which computer and bioethics might have shared concerns. The emerging studies into wet ALife are likely to provide strong empirical evidence for ALife’s most challenging hypothesis: that life is a certain set of computable functions that can be duplicated in any medium. I believe this will propel ALife into the midst of the mother of all cultural battles that has been gathering around the emergence of biotechnology. Philosophers need to pay close attention to this debate and can serve a vital role in clarifying and resolving the dispute. But even if ALife is merely a computer modeling technique that sheds light on living systems, it still has a number of significant ethical implications such as its use in the modeling of moral and ethical systems, as well as in the creation of artificial moral agents. (shrink)

Many authors have proposed constraining the behaviour of intelligent systems with ‘machine ethics’ to ensure positive social outcomes from the development of such systems. This paper critically analyses the prospects for machine ethics, identifying several inherent limitations. While machine ethics may increase the probability of ethical behaviour in some situations, it cannot guarantee it due to the nature of ethics, the computational limitations of computational agents and the complexity of the world. In addition, machine ethics, even if it were to (...) be ‘solved’ at a technical level, would be insufficient to ensure positive social outcomes from intelligent systems. (shrink)

The concept of morality underpins the moral responsibility that not only depends on the outward practices (or ‘output’, in the case of humanoid robots) of the agents but on the internal attitudes (‘input’) that rational and responsible intentioned beings generate. The primary question that has initiated extensive debate, i.e. ‘Can humanoid robots be moral?’, stems from the normative outlook where morality includes human conscience and socio-linguistic background. This paper advances the thesis that the conceptions of morality and creativity interplay with (...) linguistic human beings instead of non-linguistic humanoid robots, as humanoid robots are indeed docile automata that cannot be responsible for their actions. To eradicate human ethics in order to make way for humanoid robot ethics highlights the moral actions and adequacy that hinges the myth of creative agency and self-dependency, which a humanoid robot can scarcely express. (shrink)

This Article explores the possibility of having ethical artificial intelligence. It argues that we face a dilemma in trying to develop artificial intelligence that is ethical: either we have to be able to codify ethics as a set of rules or we have to value a machine’s ability to make ethical mistakes so that it can learn ethics like children do. Neither path seems very promising, though perhaps by thinking about the difficulties with each we may come to a better (...) understanding of artificial intelligence and ourselves. (shrink)

Suppose that an autonomous vehicle encounters a situation where (i) imposing a risk of harm on at least one person is unavoidable; and (ii) a choice about how to allocate risks of harm between different persons is required. What does morality require in these cases? Derek Leben defends a Rawlsian answer to this question. I argue that we have reason to reject Leben’s answer.

Programming computers to engage in moral reasoning is not a new idea (Anderson and Anderson 2011a). Work on the subject has yielded concrete examples of computable linguistic structures for a moral grammar (Mikhail 2007), the ethical governor architecture for autonomous weapon systems (Arkin 2009), rule-based systems that implement deontological principles (Anderson and Anderson 2011b), systems that implement utilitarian principles, and a hybrid approach to programming ethical machines (Wallach and Allen 2008). This chapter considers two philosophically informed strategies for engineering software (...) that can engage in moral reasoning: algorithms based on philosophical moral theories and analogical reasoning from standard cases.1 Based on the challenges presented to the algorithmic approach, I argue that a combination of these two strategies holds the most promise and show concrete examples of how such an architecture could be built using contemporary engineering techniques. (shrink)

The theory and philosophy of artificial intelligence has come to a crucial point where the agenda for the forthcoming years is in the air. This special volume of Minds and Machines presents leading invited papers from a conference on the “Philosophy and Theory of Artificial Intelligence” that was held in October 2011 in Thessaloniki. Artificial Intelligence is perhaps unique among engineering subjects in that it has raised very basic questions about the nature of computing, perception, reasoning, learning, language, action, interaction, (...) consciousness, humankind, life etc. etc. – and at the same time it has contributed substantially to answering these questions. There is thus a substantial tradition of work, both on AI by philosophers and of theory within AI itself. - The volume contains papers by Bostrom, Dreyfus, Gomila, O'Regan and Shagrir. (shrink)

This paper discusses different approaches incognitive science and artificial intelligenceresearch from the perspective of radicalconstructivism, addressing especially theirrelation to the biologically based theories ofvon Uexküll, Piaget as well as Maturana andVarela. In particular recent work in New AI and adaptive robotics on situated and embodiedintelligence is examined, and we discuss indetail the role of constructive processes asthe basis of situatedness in both robots andliving organisms.

Many industry leaders and academics from the field of machine ethics would have us believe that the inevitability of robots coming to have a larger role in our lives demands that robots be endowed with moral reasoning capabilities. Robots endowed in this way may be referred to as artificial moral agents. Reasons often given for developing AMAs are: the prevention of harm, the necessity for public trust, the prevention of immoral use, such machines are better moral reasoners than humans, and (...) building these machines would lead to a better understanding of human morality. Although some scholars have challenged the very initiative to develop AMAs, what is currently missing from the debate is a closer examination of the reasons offered by machine ethicists to justify the development of AMAs. This closer examination is especially needed because of the amount of funding currently being allocated to the development of AMAs coupled with the amount of attention researchers and industry leaders receive in the media for their efforts in this direction. The stakes in this debate are high because moral robots would make demands on society; answers to a host of pending questions about what counts as an AMA and whether they are morally responsible for their behavior or not. This paper shifts the burden of proof back to the machine ethicists demanding that they give good reasons to build AMAs. The paper argues that until this is done, the development of commercially available AMAs should not proceed further. (shrink)

This article focuses on a particular issue under machine ethics—that is, the nature of Artificial Moral Agents. Machine ethics is a branch of artificial intelligence that looks into the moral status of artificial agents. Artificial moral agents, on the other hand, are artificial autonomous agents that possess moral value, as well as certain rights and responsibilities. This paper demonstrates that attempts to fully develop a theory that could possibly account for the nature of Artificial Moral Agents may consider certain philosophical (...) ideas, like the standard characterizations of agency, rational agency, moral agency, and artificial agency. At the very least, the said philosophical concepts may be treated as signposts for further research on how to truly account for the nature of Artificial Moral Agents. (shrink)

What we have learnt in the last six or seven decades about virtual machinery, as a result of a great deal of science and technology, enables us to offer Darwin a new defence against critics who argued that only physical form, not mental capabilities and consciousness could be products of evolution by natural selection. The defence compares the mental phenomena mentioned by Darwin’s opponents with contents of virtual machinery in computing systems. Objects, states, events, and processes in virtual machinery which (...) we have only recently learnt how to design and build, and could not even have been thought about in Darwin’s time, can interact with the physical machinery in which they are implemented, without being identical with their physical implementation, nor mere aggregates of physical structures and processes. The existence of various kinds of virtual machinery (including both “platform” virtual machines that can host other virtual machines, e.g. operating systems, and “application” virtual machines, e.g. spelling checkers, and computer games) depends on complex webs of causal connections involving hardware and software structures, events and processes, where the specification of such causal webs requires concepts that cannot be defined in terms of concepts of the physical sciences. That indefinability, plus the possibility of various kinds of self-monitoring within virtual machinery, seems to explain some of the allegedly mysterious and irreducible features of consciousness that motivated Darwin’s critics and also more recent philosophers criticising AI. There are consequences for philosophy, psychology, neuroscience and robotics. (shrink)

Viewed from an evolutionary standpoint, emotions can be understood as situation-specific patterns of energy consumption related to behaviors that have been selected by evolution for their survival value, such as environmental exploration, flight or fight, and socialization. In the present article, the energy linked with emotions is investigated by a strictly energy-based simulation of the evolution of simple autonomous agents provided with random cognitive and motor capacities and operating among food and predators. Emotions are translated into evolving patterns of energy (...) consumption related to situation-specific behaviors. As a result, a variety of behaviors resembling emotions emerge, each with a specific pattern of energy consumption. There is little difference between emotion-like behaviors on individual and collective levels. The fact that these patterns evolve under a strictly energetic selection regime indicates the interesting role and adaptive advantage of emotions as situation-specific energy savers. Although gained through an approach characterized by a highly reduced complexity, these findings support a quantitative understanding of emotions that may improve the existing qualitative approaches and could have far-reaching implications for further research on and conceptualization of emotions and their interactions with cognition. (shrink)

This paper focuses on the research field of machine ethics and how it relates to a technological singularity—a hypothesized, futuristic event where artificial machines will have greater-than-human-level intelligence. One problem related to the singularity centers on the issue of whether human values and norms would survive such an event. To somehow ensure this, a number of artificial intelligence researchers have opted to focus on the development of artificial moral agents, which refers to machines capable of moral reasoning, judgment, and decision-making. (...) To date, different frameworks on how to arrive at these agents have been put forward. However, there seems to be no hard consensus as to which framework would likely yield a positive result. With the body of work that they have contributed in the study of moral agency, philosophers may contribute to the growing literature on artificial moral agency. While doing so, they could also think about how the said concept could affect other important philosophical concepts. (shrink)

Humans are becoming increasingly dependent on the ‘say-so' of machines, such as computers, smartphones, and robots. In epistemology, knowledge based on what you have been told is called ‘testimony' and being able to give and receive testimony is a prerequisite for engaging in many social roles. Should robots and other autonomous intelligent machines be considered epistemic testifiers akin to those of humans? This chapter attempts to answer this question as well as explore the implications of robot testimony for the criminal (...) justice system. Few are in agreement as to the ‘types' of agents that can provide testimony. The chapter surveys three well-known approaches and shows that on two of these approaches being able to provide testimony is bound up with the possession of intentional mental states. Through a discussion of computational and folk-psychological approaches to intentionality, it is argued that a good case can be made for robots fulfilling all three definitions. (shrink)

The advent of the intelligent robot has occupied a significant position in society over the past decades and has given rise to new issues in society. As we know, the primary aim of artificial intelligence or robotic research is not only to develop advanced programs to solve our problems but also to reproduce mental qualities in machines. The critical claim of artificial intelligence advocates is that there is no distinction between mind and machines and thus they argue that there are (...) possibilities for machine ethics, just as human ethics. Unlike computer ethics, which has traditionally focused on ethical issues surrounding human use of machines, AI or machine ethics is concerned with the behaviour of machines towards human users and perhaps other machines as well, and the ethicality of these interactions. The ultimate goal of machine ethics, according to the AI scientists, is to create a machine that itself follows an ideal ethical principle or a set of principles; that is to say, it is guided by this principle or these principles in decisions it makes about possible courses of action it could take. Thus, machine ethics task of ensuring ethical behaviour of an artificial agent. Although, there are many philosophical issues related to artificial intelligence, but our attempt in this paper is to discuss, first, whether ethics is the sort of thing that can be computed. Second, if we are ascribing mind to machines, it gives rise to ethical issues regarding machines. And if we are not drawing the difference between mind and machines, we are not only redefining specifically human mind but also the society as a whole. Having a mind is, among other things, having the capacity to make voluntary decisions and actions. The notion of mind is central to our ethical thinking, and this is because the human mind is self-conscious, and this is a property that machines lack, as yet. (shrink)

In this paper, we present a formal interpretive account of Jean Piaget’s theory of the morality that regulates social exchanges, which we call interactional morality. First, we place Piaget’s conception in the context of his epistemological and sociological works. Then, we review the core of that conception: the two types of interactional moralities that Piaget identified to be usual in social exchanges, and the role that the notion of respect-for-the-other plays in their definition. Next, we analyze the main features of (...) social exchanges that are subject to each of those two types of interactional moralities. Following, we formalize Piaget’s account of how the two types of interactional moralities can be put to regulate social exchanges in which values, norms, beliefs, and goals are exchanged, so that interactional commitments can be established among the participants. Then, we sketch a moral reasoning system able to support formal reasonings about interactional moralities. Finally, we illustrate the usefulness of the overall formal model of interactional moralities that resulted from our study by applying it to the explication of some of the conclusions achieved by Piaget in his moral studies about children games. (shrink)

The development of artificial intelligence will require systems of ethical decision making to be adapted for automatic computation. However, projects to implement moral reasoning in artificial moral agents so far have failed to satisfactorily address the widespread disagreement between competing approaches to moral philosophy. In this paper I argue that the proper response to this situation is to design machines to be fundamentally uncertain about morality. I describe a computational framework for doing so and show that it efficiently resolves common (...) obstacles to the implementation of moral philosophy in intelligent machines. (shrink)

In this book, the extend to which we can equip artificial agents with moral reasoning capacity is investigated. Attempting to create artificial agents with moral reasoning capabilities challenges our understanding of morality and moral reasoning to its utmost. It also helps philosophers dealing with the inherent complexity of modern organizations. Modern society with large multi-national organizations and extensive information infrastructures provides a backdrop for moral theories that is hard to encompass through mere theorising. Computerized support for theorising is needed to (...) be able to fully grasp and address the inherent complexity. Using moral reasoning capacity will help us addressing the challenges that technological artefacts pose. They do not only contain information about us, they start to act on our behalves.With the increasing autonomy comes an increased need to ensure that their behaviour is in line with what we expect from them. To investigate and address these issues a laboratoy for philosophy is outlined: SophoLab. It consists of a methodology; a framework of modal logic, DEAL; and multi-agent software systems. SophoLab provides the basis for an experimental, computational philosophy. Its viability and usefulness are demonstrated through several experiments. (shrink)

Designing in Ethics provides a compilation of well-curated essays that tackle the ethical issues that surround technological design and argue that ethics must form a constitutive part of the designing process and a foundation in our institutions and practices. The appropriation of a design approach to applied ethics is argued as a means by which ethical issues that implicate technological artifact may be achieved.

This article argues that existing approaches to programming ethical AI fail to resolve a serious moral-semantic trilemma, generating interpretations of ethical requirements that are either too semantically strict, too semantically flexible, or overly unpredictable. This paper then illustrates the trilemma utilizing a recently proposed ‘general ethical dilemma analyzer,’ _GenEth_. Finally, it uses empirical evidence to argue that human beings resolve the semantic trilemma using general cognitive and motivational processes involving ‘mental time-travel,’ whereby we simulate different possible pasts and futures. I (...) demonstrate how mental time-travel psychology leads us to resolve the semantic trilemma through a six-step process of interpersonal negotiation and renegotiation, and then conclude by showing how comparative advantages in processing power would plausibly cause AI to use similar processes to solve the semantic trilemma more reliably than we do, leading AI to make better moral-semantic choices than humans do by our very own lights. (shrink)

A properly programmed artificially intelligent agent may eventually have one duty, the duty to satisfice expected welfare. We explain this claim and defend it against objections.

The aim of this paper is to offer an analysis of the notion of artificial moral agent (AMA) and of its impact on human beings’ self-understanding as moral agents. Firstly, I introduce the topic by presenting what I call the Continuity Approach. Its main claim holds that AMAs and human moral agents exhibit no significant qualitative difference and, therefore, should be considered homogeneous entities. Secondly, I focus on the consequences this approach leads to. In order to do this I take (...) into consideration the work of Bostrom and Dietrich, who have radically assumed this viewpoint and thoroughly explored its implications. Thirdly, I present an alternative approach to AMAs—the Discontinuity Approach—which underscores an essential difference between human moral agents and AMAs by tackling the matter from another angle. In this section I concentrate on the work of Johnson and Bryson and I highlight the link between their claims and Heidegger’s and Jonas’s sug-gestions concerning the relationship between human beings and technological products. In conclusion I argue that, although the Continuity Approach turns out to be a necessary postulate to the machine ethics project, the Discontinuity Approach highlights a relevant distinction between AMAs and human moral agents. On this account, the Discontinuity Approach generates a clearer understanding of what AMAs are, of how we should face the moral issues they pose, and, finally, of the difference that separates machine ethics from moral philosophy. (shrink)

The frame problem is the difficulty of explaining how non-magical systems think and act in ways that are adaptively sensitive to context-dependent relevance. Influenced centrally by Heideggerian phenomenology, Hubert Dreyfus has argued that the frame problem is, in part, a consequence of the assumption (made by mainstream cognitive science and artificial intelligence) that intelligent behaviour is representation-guided behaviour. Dreyfus' Heideggerian analysis suggests that the frame problem dissolves if we reject representationalism about intelligence and recognize that human agents realize the property (...) of thrownness (the property of being always already embedded in a context). I argue that this positive proposal is incomplete until we understand exactly how the properties in question may be instantiated in machines like us. So, working within a broadly Heideggerian conceptual framework, I pursue the character of a representation-shunning thrown machine. As part of this analysis, I suggest that the frame problem is, in truth, a two-headed beast. The intra-context frame problem challenges us to say how a purely mechanistic system may achieve appropriate, flexible and fluid action within a context. The inter-context frame problem challenges us to say how a purely mechanistic system may achieve appropriate, flexible and fluid action in worlds in which adaptation to new contexts is open-ended and in which the number of potential contexts is indeterminate. Drawing on the field of situated robotics, I suggest that the intra-context frame problem may be neutralized by systems of special-purpose adaptive couplings, while the inter-context frame problem may be neutralized by systems that exhibit the phenomenon of continuous reciprocal causation. I also defend the view that while continuous reciprocal causation is in conflict with representational explanation, special-purpose adaptive coupling, as well as its associated agential phenomenology, may feature representations. My proposal has been criticized recently by Dreyfus, who accuses me of propagating a cognitivist misreading of Heidegger, one that, because it maintains a role for representation, leads me seriously astray in my handling of the frame problem. I close by responding to Dreyfus' concerns. (shrink)

Better instruments to predict the future evolution of artificial intelligence (AI) are needed, as the destiny of our civilization depends on it. One of the ways to such prediction is the analysis of the convergent drives of any future AI, started by Omohundro. We show that one of the convergent drives of AI is a militarization drive, arising from AI’s need to wage a war against its potential rivals by either physical or software means, or to increase its bargaining power. (...) This militarization trend increases global catastrophic risk or even existential risk during AI takeoff, which includes the use of nuclear weapons against rival AIs, blackmail by the threat of creating a global catastrophe, and the consequences of a war between two AIs. As a result, even benevolent AI may evolve into potentially dangerous military AI. The type and intensity of militarization drive depend on the relative speed of the AI takeoff and the number of potential rivals. We show that AI militarization drive and evolution of national defense will merge, as a superintelligence created in the defense environment will have quicker takeoff speeds, but a distorted value system. We conclude with peaceful alternatives. (shrink)

We argue that, to be trustworthy, Computa- tional Intelligence (CI) has to do what it is entrusted to do for permissible reasons and to be able to give rationalizing explanations of its behavior which are accurate and gras- pable. We support this claim by drawing par- allels with trustworthy human persons, and we show what difference this makes in a hypo- thetical CI hiring system. Finally, we point out two challenges for trustworthy CI and sketch a mechanism which could be (...) used to gener- ate sufficiently accurate as well as graspable rationalizing explanations for CI behavior. (shrink)

To create fair and accountable AI and robotics, we need precise regulation and better methods to certify, explain, and audit inscrutable systems.

In this paper, we take moral agency to be that context in which a particular agent can, appropriately, be held responsible for her actions and their consequences. In order to understand moral agency, we will discuss what it would take for an artifact to be a moral agent. For reasons that will become clear over the course of the paper, we take the artifactual question to be a useful way into discussion but ultimately misleading. We set out a number of (...) conceptual pre-conditions for being a moral agent and then outline how one should — and should not — go about attributing moral agency. In place of a litmus test for such an agency — such as Allen et al.'s Moral Turing Test — we suggest some tools from the conceptual spaces theory and the unified conceptual space theory for mapping out the nature and extent of that agency. (shrink)