Our ability to locate moral responsibility is often thought to be a necessary condition for conducting morally permissible medical practice, engaging in a just war, and other high-stakes endeavors. Yet, with increasing reliance upon artificially intelligent systems, we may be facing a wideningresponsibility gap, which, some argue, cannot be bridged by traditional concepts of responsibility. How then, if at all, can we make use of crucial emerging technologies? According to Colin Allen and Wendell Wallach, the advent of so-called ‘artificial moral (...) agents’ (AMAs) is inevitable. Still, this notion may seem to push back the problem, leaving those who have an interest in developing autonomous technology with a dilemma. We may need to scale-back our efforts at deploying AMAs (or at least maintain human oversight); otherwise, we must rapidly and drastically update our moral and legal norms in a way that ensures responsibility for potentially avoidable harms. This paper invokes contemporary accounts of responsibility in order to show how artificially intelligent systems might be held responsible. Although many theorists are concerned enough to develop artificial conceptions ofagencyor to exploit our present inability to regulate valuable innovations, the proposal here highlights the importance of—and outlines a plausible foundation for—a workable notion ofartificial moral responsibility. (shrink)

As inductive decision-making procedures, the inferences made by machine learning programs are subject to underdetermination by evidence and bear inductive risk. One strategy for overcoming these challenges is guided by a presumption in philosophy of science that inductive inferences can and should be value-free. Applied to machine learning programs, the strategy assumes that the influence of values is restricted to data and decision outcomes, thereby omitting internal value-laden design choice points. In this paper, I apply arguments from feminist philosophy of (...) science to machine learning programs to make the case that the resources required to respond to these inductive challenges render critical aspects of their design constitutively value-laden. I demonstrate these points specifically in the case of recidivism algorithms, arguing that contemporary debates concerning fairness in criminal justice risk-assessment programs are best understood as iterations of traditional arguments from inductive risk and demarcation, and thereby establish the value-laden nature of automated decision-making programs. Finally, in light of these points, I address opportunities for relocating the value-free ideal in machine learning and the limitations that accompany them. (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)

The ability to interact in a natural language profoundly changes devices’ interfaces and potential applications of speaking technologies. Concurrently, this phenomenon challenges our mainstream theories of knowledge, such as how to analyze linguistic outputs of devices under existing anthropocentric theoretical assumptions. In section 1, I present the topic of machines that speak, connecting between Descartes and Generative AI. In section 2, I argue that accepted testimonial theories of knowledge and justification commonly reject the possibility that a speaking technological artifact can (...) give testimony. In section 3, I identify three assumptions underlying the view that rejects conversational AIs – AI-based technologies that converse, as testifiers: conversational AIs (1) lack intentions, (2) cannot be normatively assessed, and (3) cannot constitute an object in trust relations, while humans can. In section 4, I propose the concept ‘AI-testimony’ for analyzing outputs of conversational AIs, suggesting three conditions for technologies to deliver AI-testimony: (1) content is propositional, (2) generated and delivered with no other human directly involved, (3) the output is perceived as phenomenologically similar to that of a human. I conclude that this concept overcomes the limitations of the anthropocentric concept of testimony, opening future directions of research without associating conversational AIs with human-like agency. (shrink)

In this paper we explore whether or not the use of risk assessment tools in criminal sentencing can be made compatible with a retributivist justification of punishment. While there has been considerable discussion of the accuracy and fairness of these tools, such discussion assumes that one’s recidivism risk is relevant to the severity of punishment that one should receive. But this assumption only holds on certain accounts of punishment, and seems to conflict with retributivist justifications of punishment. Drawing on the (...) broader desert literature, we explain the source of this conflict, and suggest that a retributivist approach on which the severity of punishment partly depends on one’s character in addition to their acts offers some hope of reconciling retributivism with the use of risk assessment tools in considerations of sentencing reduction in particular. Ultimately, however, even this limited attempt at reconciliation fails, so long as risk assessment tools fail to distinguish between risk that one is responsible for, and risk that one is not responsible for. (shrink)

We argue why interpretability should have primacy alongside empiricism for several reasons: first, if machine learning models are beginning to render some of the high-risk healthcare decisions instead of clinicians, these models pose a novel medicolegal and ethical frontier that is incompletely addressed by current methods of appraising medical interventions like pharmacological therapies; second, a number of judicial precedents underpinning medical liability and negligence are compromised when ‘autonomous’ ML recommendations are considered to be en par with human instruction in specific (...) contexts; third, explainable algorithms may be more amenable to the ascertainment and minimisation of biases, with repercussions for racial equity as well as scientific reproducibility and generalisability. We conclude with some reasons for the ineludible importance of interpretability, such as the establishment of trust, in overcoming perhaps the most difficult challenge ML will face in a high-stakes environment like healthcare: professional and public acceptance. (shrink)

The challenge of designing computer systems and robots with the ability to make moral judgments is stepping out of science fiction and moving into the laboratory. Engineers and scholars, anticipating practical necessities, are writing articles, participating in conference workshops, and initiating a few experiments directed at substantiating rudimentary moral reasoning in hardware and software. The subject has been designated by several names, including machine ethics, machine morality, artificial morality, or computational morality. Most references to the challenge elucidate one facet or (...) another of what is a very rich topic. This paper will offer a brief overview of the many dimensions of this new field of inquiry. (shrink)

Anonymity is a form of nonidentifiability which I define as noncoordinatability of traits in a given respect. This definition broadens the concept, freeing it from its primary association with naming. I analyze different ways anonymity can be realized. I also discuss some ethical issues, such as privacy, accountability and other values which anonymity may serve or undermine. My theory can also conceptualize anonymity in information systems where, for example, privacy and accountability are at issue.

This paper investigates how the introduction of AI to decision making increases moral distance and recommends the ethics of care to augment the ethical examination of AI decision making. With AI decision making, face-to-face interactions are minimized, and decisions are part of a more opaque process that humans do not always understand. Within decision-making research, the concept of moral distance is used to explain why individuals behave unethically towards those who are not seen. Moral distance abstracts those who are impacted (...) by the decision and leads to less ethical decisions. The goal of this paper is to identify and analyze the moral distance created by AI through both proximity distance (in space, time, and culture) and bureaucratic distance (derived from hierarchy, complex processes, and principlism). We then propose the ethics of care as a moral framework to analyze the moral implications of AI. The ethics of care brings to the forefront circumstances and context, interdependence, and vulnerability in analyzing algorithmic decision making. (shrink)

In some situations in which undesirable collective effects occur, it is very hard, if not impossible, to hold any individual reasonably responsible. Such a situation may be referred to as the problem of many hands. In this paper we investigate how the problem of many hands can best be understood and why, and when, it exactly constitutes a problem. After analyzing climate change as an example, we propose to define the problem of many hands as the occurrence of a gap (...) in the distribution of responsibility that may be considered morally problematic. Whether a gap is morally problematic, we suggest, depends on the reasons why responsibility is distributed. This, in turn, depends, at least in part, on the sense of responsibility employed, a main distinction being that between backward-looking and forward-looking responsibility. (shrink)

Artificial intelligence and robotic technologies have become nearly ubiquitous. In some ways, the developments have likely helped us, but in other ways sophisticated technologies set back our interests. Among the latter sort is what has been dubbed the ‘severance problem’—the idea that technologies sever our connection to the world, a connection which is necessary for us to flourish and live meaningful lives. I grant that the severance problem is a threat we should mitigate and I ask: how can we stave (...) it off? In particular, the fact that some technologies exhibit behavior that is unclear to us seems to constitute a kind of severance. Building upon contemporary work on moral responsibility, I argue for a mechanism I refer to as ‘technological answerability’, namely the capacity to recognize human demands for answers and to respond accordingly. By designing select devices—such as robotic assistants and personal AI programs—for increased answerability, we see at least one way of satisfying our demands for answers and thereby retaining our connection to a world increasingly occupied by technology. (shrink)

Machine Learning (ML) and Artificial Intelligence (AI) are for many the defining features of the early twenty-first century. With such a provocation, this essay considers how one might understand the relational philosophies articulated by Indigenous learning scientists, Indigenous technologists and feminine philosophers of education as co-constitutive of an ensemble mediating or regulating an educative philosophy interfacing with ML/AI. In these mediations, differing vocabularies – kin, the one caring, cooperative – are recognized for their ethical commitments, yet challenging epistemic claims in (...) the contexts of ML. Similarly, ML poses some questions to claims made about relational and Indigenous epistemologies, where the latter is perceived as separated from and unaffected by computation specifically or algorithmic societies generally. This essay seeks to gain several vantage points to explore and complicate how diverse relational philosophies can address ML and perhaps reconsider their own critical practices. (shrink)

A human driver and an automated driving system might share control of automated vehicles in the near future. This raises many concerns associated with the assignment of responsibility for negative outcomes caused by them; one is that the human driver might be required to bear the brunt of moral and legal responsibilities. The psychological consequences of responsibility misattribution have not yet been examined. We designed a hypothetical crash similar to Uber’s 2018 fatal crash. We incorporated five legal responsibility attributions. Participants (...) chose their preferred liability attribution and then were randomly assigned into one of the five actual liability attribution conditions. They then responded to a series of questions concerning liability assignment, the crash, and AVs. Slightly more than 50% of participants thought that the human driver should bear full or primary liability. Legal responsibility misattribution negatively influenced these mentioned responses, regardless of overly attributing human or manufacturer liability. Overly attributing human liability had more negative influences. Improper liability attribution might hinder the adoption of AVs. Public opinion should not be ignored in developing a legal framework for AVs. (shrink)

As the aim of the responsible robotics initiative is to ensure that responsible practices are inculcated within each stage of design, development and use, this impetus is undergirded by the alignment of ethical and legal considerations towards socially beneficial ends. While every effort should be expended to ensure that issues of responsibility are addressed at each stage of technological progression, irresponsibility is inherent within the nature of robotics technologies from a theoretical perspective that threatens to thwart the endeavour. This is (...) because the concept of responsibility, despite being treated as such, is not monolithic: rather this seemingly unified concept consists of converging and confluent concepts that shape the idea of what we colloquially call responsibility. From a different perspective, robotics will be simultaneously responsible and irresponsible depending on the particular concept of responsibility that is foregrounded: an observation that cuts against the grain of the drive towards responsible robotics. This problem is further compounded by responsible design and development as contrasted to responsible use. From a different perspective, the difficulty in defining the concept of responsibility in robotics is because human responsibility is the main frame of reference. Robotic systems are increasingly expected to achieve the human-level performance, including the capacities associated with responsibility and other criteria which are necessary to act responsibly. This subsists within a larger phenomenon where the difference between humans and non-humans, be it animals or artificial systems, appears to be increasingly blurred, thereby disrupting orthodox understandings of responsibility. This paper seeks to supplement the responsible robotics impulse by proposing a complementary set of human rights directed specifically against the harms arising from robotic and artificial intelligence technologies. The relationship between responsibilities of the agent and the rights of the patient suggest that a rights regime is the other side of responsibility coin. The major distinction of this approach is to invert the power relationship: while human agents are perceived to control robotic patients, the prospect for this to become reversed is beginning. As robotic technologies become ever more sophisticated, and even genuinely complex, asserting human rights directly against robotic harms become increasingly important. Such an approach includes not only developing human rights that ‘protect’ humans but also ‘strengthen’ people against the challenges introduced by robotics and AI [This distinction parallels Berlin’s negative and positive concepts of liberty ], by emphasising the social and reflective character of the notion of humanness as well as the difference between the human and nonhuman. This will allow using the human frame of reference as constitutive of, rather than only subject to, the robotic and AI technologies, where it is human and not technology characteristics that shape the human rights framework in the first place. (shrink)

This contribution deals with the ethical challenges arising from the IoT landscape with reference to a specific context, i.e. the realm of agri-food. In this sector, innumerable web-connected tools, platforms and sensors are constantly interacting with consumers/users/citizens, by reshaping and redefining the core elements and functions of machine–human being relationships. By sketching out the main pillars which ethics of the Internet of Food is founded on, my argument posits that the civic hybridization of knowledge production mediated by IoT technologies may (...) create breeding ground for the move towards an ‘ethical in-design’ approach to the IoF-driven smart systems. (shrink)

Decision-making assisted by algorithms developed by machine learning is increasingly determining our lives. Unfortunately, full opacity about the process is the norm. Would transparency contribute to restoring accountability for such systems as is often maintained? Several objections to full transparency are examined: the loss of privacy when datasets become public, the perverse effects of disclosure of the very algorithms themselves (“gaming the system” in particular), the potential loss of companies’ competitive edge, and the limited gains in answerability to be expected (...) since sophisticated algorithms usually are inherently opaque. It is concluded that, at least presently, full transparency for oversight bodies alone is the only feasible option; extending it to the public at large is normally not advisable. Moreover, it is argued that algorithmic decisions preferably should become more understandable; to that effect, the models of machine learning to be employed should either be interpreted ex post or be interpretable by design ex ante. (shrink)

A self-fulfilling prophecy is, roughly, a prediction that brings about its own truth. Although true predictions are hard to fault, self-fulfilling prophecies are often regarded with suspicion. In this article, we vindicate this suspicion by explaining what self-fulfilling prophecies are and what is problematic about them, paying special attention to how their problems are exacerbated through automated prediction. Our descriptive account of self-fulfilling prophecies articulates the four elements that define them. Based on this account, we begin our critique by showing (...) that typical self-fulfilling prophecies arise due to mistakes about the relationship between a prediction and its object. Such mistakes—along with other mistakes in predicting or in the larger practical endeavor—are easily overlooked when the predictions turn out true. Thus we note that self-fulfilling prophecies prompt no error signals; truth shrouds their mistakes from humans and machines alike. Consequently, self-fulfilling prophecies create several obstacles to accountability for the outcomes they produce. We conclude our critique by showing how failures of accountability, and the associated failures to make corrections, explain the connection between self-fulfilling prophecies and feedback loops. By analyzing the complex relationships between accuracy and other evaluatively significant features of predictions, this article sheds light both on the special case of self-fulfilling prophecies and on the ethics of prediction more generally. (shrink)

Background Big data and AI applications now play a major role in many health contexts. Much research has already been conducted on ethical and social challenges associated with these technologies. Likewise, there are already some studies that investigate empirically which values and attitudes play a role in connection with their design and implementation. What is still in its infancy, however, is the comparative investigation of the perspectives of different stakeholders. Methods To explore this issue in a multi-faceted manner, we conducted (...) semi-structured interviews as well as focus group discussions with patients and clinicians. These empirical methods were used to gather interviewee’s views on the opportunities and challenges of medical AI and other data-intensive applications. Results Different clinician and patient groups are exposed to medical AI to differing degrees. Interviewees expect and demand that the purposes of data processing accord with patient preferences, and that data are put to effective use to generate social value. One central result is the shared tendency of clinicians and patients to maintain individualistic ascriptions of responsibility for clinical outcomes. Conclusions Medical AI and the proliferation of data with import for health-related inferences shape and partially reconfigure stakeholder expectations of how these technologies relate to the decision-making of human agents. Intuitions about individual responsibility for clinical outcomes could eventually be disrupted by the increasing sophistication of data-intensive and AI-driven clinical tools. Besides individual responsibility, systemic governance will be key to promote alignment with stakeholder expectations in AI-driven and data-intensive health settings. (shrink)

Public entities around the world are increasingly deploying artificial intelligence and algorithmic decision-making systems to provide public services or to use their enforcement powers. The rationale for the public sector to use these systems is similar to private sector: increase efficiency and speed of transactions and lower the costs. However, public entities are first and foremost established to meet the needs of the members of society and protect the safety, fundamental rights, and wellbeing of those they serve. Currently AI systems (...) are deployed by the public sector at various administrative levels without robust due diligence, monitoring, or transparency. This paper critically maps out the challenges in procurement of AI systems by public entities and the long-term implications necessitating AI-specific procurement guidelines and processes. This dual-prong exploration includes the new complexities and risks introduced by AI systems, and the institutional capabilities impacting the decision-making process. AI-specific public procurement guidelines are urgently needed to protect fundamental rights and due process. (shrink)

The crash of two 737 MAX passenger aircraft in late 2018 and early 2019, and subsequent grounding of the entire fleet of 737 MAX jets, turned a global spotlight on Boeing’s practices and culture. Explanations for the crashes include: design flaws within the MAX’s new flight control software system designed to prevent stalls; internal pressure to keep pace with Boeing’s chief competitor, Airbus; Boeing’s lack of transparency about the new software; and the lack of adequate monitoring of Boeing by the (...) FAA, especially during the certification of the MAX and following the first crash. While these and other factors have been the subject of numerous government reports and investigative journalism articles, little to date has been written on the ethical significance of the accidents, in particular the ethical responsibilities of the engineers at Boeing and the FAA involved in designing and certifying the MAX. Lessons learned from this case include the need to strengthen the voice of engineers within large organizations. There is also the need for greater involvement of professional engineering societies in ethics-related activities and for broader focus on moral courage in engineering ethics education. (shrink)

This article presents a conceptual investigation into the value impacts and relations of algorithms in the domain of justice and security. As a conceptual investigation, it represents one step in a value sensitive design based methodology. Here, we explicate and analyse the expression of values of accuracy, privacy, fairness and equality, property and ownership, and accountability and transparency in this context. We find that values are sensitive to disvalue if algorithms are designed, implemented or deployed inappropriately or without sufficient consideration (...) for their value impacts, potentially resulting in problems including discrimination and constrained autonomy. Furthermore, we outline a framework of conceptual relations of values indicated by our analysis, and potential value tensions in their implementation and deployment with a view towards supporting future research, and supporting the value sensitive design of algorithms in justice and security. (shrink)

The task of this essay is to respond to the question concerning robots and responsibility—to answer for the way that we understand, debate, and decide who or what is able to answer for decisions and actions undertaken by increasingly interactive, autonomous, and sociable mechanisms. The analysis proceeds through three steps or movements. It begins by critically examining the instrumental theory of technology, which determines the way one typically deals with and responds to the question of responsibility when it involves technology. (...) It then considers three instances where recent innovations in robotics challenge this standard operating procedure by opening gaps in the usual way of assigning responsibility. The innovations considered in this section include: autonomous technology, machine learning, and social robots. The essay concludes by evaluating the three different responses—instrumentalism 2.0, machine ethics, and hybrid responsibility—that have been made in face of these difficulties in an effort to map out the opportunities and challenges of and for responsible robotics. (shrink)

In light of the potential long-term societal and economic benefits of novel nano-enabled products, there is an evident need for research and development to focus on closing the gap in nano-materials safety. Concurrent reflection on the impact of decision-making tools, which may lack the capability to assist sophisticated judgements around the risks and benefits of the introduction of novel products, is essential. This paper addresses the potential for extant decision support tools to default to a precautionary principle position in the (...) face of uncertainty. A more utilitarian-based approach could be facilitated by adding simple methods to formulate realistic hypotheses, which would assist non-specialists to make more nuanced decisions in terms of managing the risks of introducing new NMs. A decision support analytical framework is applied to identify the potential risks and benefits of novel nano-enabled products such as textiles with in-built enhanced antimicrobial activity for the prevention of nosocomial infections produced by spray or sonochemical coating possesses. While the results demonstrate valuable societal and environmental benefits compared to conventional products, due to uncertainty regarding the possible hazard to humans, sizable risks were identified in some cases due to the precautionary principle. (shrink)

In this article, we develop the concept of Transparency by Design that serves as practical guidance in helping promote the beneficial functions of transparency while mitigating its challenges in automated-decision making environments. With the rise of artificial intelligence and the ability of AI systems to make automated and self-learned decisions, a call for transparency of how such systems reach decisions has echoed within academic and policy circles. The term transparency, however, relates to multiple concepts, fulfills many functions, and holds different (...) promises that struggle to be realized in concrete applications. Indeed, the complexity of transparency for ADM shows tension between transparency as a normative ideal and its translation to practical application. To address this tension, we first conduct a review of transparency, analyzing its challenges and limitations concerning automated decision-making practices. We then look at the lessons learned from the development of Privacy by Design, as a basis for developing the Transparency by Design principles. Finally, we propose a set of nine principles to cover relevant contextual, technical, informational, and stakeholder-sensitive considerations. Transparency by Design is a model that helps organizations design transparent AI systems, by integrating these principles in a step-by-step manner and as an ex-ante value, not as an afterthought. (shrink)

The paper examines how the current evolution and growth of ICTs enables a greater number of individuals to communicate and interact with each other on a larger scale: this phenomenon enables people to cross the conventional boundaries set up across modernity. The presence of diverse barriers does not however disappear, and we therefore still experience cultural, political, legal and moral boundaries in the globalised Information Society. The paper suggests that the issue of boundaries is to be understood, primarily, in philosophical (...) terms, in order to explain in what sense the fundamental historicity and situatedness of human beings does not necessarily prevent them from being able to cross the boundaries of their own experiences and to communicate with other human beings. Our line of reasoning is based on the idea that the epistemic and moral subject is neither a pure universal subject always transcending the limits of its experience nor the mere product of the context in which it find itself embedded. By virtue of its continued existence and experience of any given context, the epistemic and moral subject is asked to respond to the constraining affordances of data and to account for the moral calls of the other. This intertwinement of responsiveness and accountability towards what is "over there" confers us an original understanding of what is "here". (shrink)

Decision-making assisted by algorithms developed by machine learning is increasingly determining our lives. Unfortunately, full opacity about the process is the norm. Would transparency contribute to restoring accountability for such systems as is often maintained? Several objections to full transparency are examined: the loss of privacy when datasets become public, the perverse effects of disclosure of the very algorithms themselves, the potential loss of companies’ competitive edge, and the limited gains in answerability to be expected since sophisticated algorithms usually are (...) inherently opaque. It is concluded that, at least presently, full transparency for oversight bodies alone is the only feasible option; extending it to the public at large is normally not advisable. Moreover, it is argued that algorithmic decisions preferably should become more understandable; to that effect, the models of machine learning to be employed should either be interpreted ex post or be interpretable by design ex ante. (shrink)

Editors’ Overview: Moral Responsibility in Technology and Engineering Content Type Journal Article Category Original Paper Pages 1-11 DOI 10.1007/s11948-011-9285-z Authors Neelke Doorn, Department of Technology, Policy and Management, Delft University of Technology, P.O. Box 5015, 2600 GA Delft, The Netherlands Ibo van de Poel, Department of Technology, Policy and Management, Delft University of Technology, P.O. Box 5015, 2600 GA Delft, The Netherlands Journal Science and Engineering Ethics Online ISSN 1471-5546 Print ISSN 1353-3452 Journal Volume Volume 18 Journal Issue Volume 18, (...) Number 1. (shrink)

How may human agents come to trust artificial agents? At present, since the trust involved is non-normative, this would seem to be a slow process, depending on the outcomes of the transactions. Some more options may soon become available though. As debated in the literature, humans may meet bots as they are embedded in an institution. If they happen to trust the institution, they will also trust them to have tried out and tested the machines in their back corridors; as (...) a consequence, they approach the robots involved as being trustworthy. Properly speaking, users rely on the overall accountability of the institution. Besides this option we explore some novel ways for trust development: trust becomes normatively laden and thereby the mechanism of exclusive reliance on the normative force of trust may come into play - the efficacy of which has already been proven for persons meeting face-to-face or over the Internet. For one thing, machines may evolve into moral machines, or machines skilled in the art of deception. While both developments might seem to facilitate proper trust and turn as-if trust into a feasible option, they are hardly to be taken seriously. For another, the new trend in robotics is towards coactivity between human and machine operators in a team. Inside the team trust is a necessity for smooth operations. In support of this, humans in particular need to be able to develop and maintain accurate mental models of their machine counterparts. Nevertheless, the trust involved is bound to remain non-normative. It is argued, though, that excellent opportunities exist to build relations of trust toward outside users who are pondering their reliance on the coactive team. The task of managing this trust has to be allotted to human operators of the team, who operate as linking pin between the outside world and the team. Since the robotic team has now been turned into an anthropomorphic team, users may well develop normative trust towards them; correspondingly, trusting the team in as-if fashion becomes feasible. (shrink)

Decision-making assisted by algorithms developed by machine learning is increasingly determining our lives. Unfortunately, full opacity about the process is the norm. Would transparency contribute to restoring accountability for such systems as is often maintained? Several objections to full transparency are examined: the loss of privacy when datasets become public, the perverse effects of disclosure of the very algorithms themselves, the potential loss of companies’ competitive edge, and the limited gains in answerability to be expected since sophisticated algorithms usually are (...) inherently opaque. It is concluded that, at least presently, full transparency for oversight bodies alone is the only feasible option; extending it to the public at large is normally not advisable. Moreover, it is argued that algorithmic decisions preferably should become more understandable; to that effect, the models of machine learning to be employed should either be interpreted ex post or be interpretable by design ex ante. (shrink)

There are many ways to avoid responsibility, for example, explaining what happens as the work of the gods, fate, society, or the system. For engineers, “technology” or “the organization” will serve this purpose quite well. We may distinguish at least nine (related) senses of “responsibility”, the most important of which are: (a) responsibility-as-causation (the storm is responsible for flooding), (b) responsibility-as-liability (he is the person responsible and will have to pay), (c) responsibility-as-competency (he’s a responsible person, that is, he’s rational), (...) (d) responsibility-as-office (he’s the responsible person, that is, the person in charge), and (e) a responsibility-as-domain-of-tasks (these are her responsibilities, that is, the things she is supposed to do). For all but the causal sense of responsibility, responsibility may be taken (in a relatively straightforward sense)—and generally is. Why then would anyone want to claim that certain technologies make it impossible to attribute responsibility to engineers (or anyone else)? In this paper, I identify seven arguments for that claim and explain why each is fallacious. The most important are: (1) the argument from “many hands”, (2) the argument from individual ignorance, and (3) the argument from blind forces. Each of these arguments makes the same fundamental mistake, the assumption that a certain factual situation, being fixed, settles responsibility, that is, that individuals, either individually or by some group decision, cannot take responsibility. I conclude by pointing out the sort of decisions (and consequences) engineers have explicitly taken responsibility for and why taking responsibility for them is rational, all things considered. There is no technological bar to such responsibility. (shrink)

Making good decisions in extremely complex and difficult processes and situations has always been both a key task as well as a challenge in the clinic and has led to a large amount of clinical, legal and ethical routines, protocols and reflections in order to guarantee fair, participatory and up-to-date pathways for clinical decision-making. Nevertheless, the complexity of processes and physical phenomena, time as well as economic constraints and not least further endeavours as well as achievements in medicine and healthcare (...) continuously raise the need to evaluate and to improve clinical decision-making. This article scrutinises if and how clinical decision-making processes are challenged by the rise of so-called artificial intelligence-driven decision support systems (AI-DSS). In a first step, this article analyses how the rise of AI-DSS will affect and transform the modes of interaction between different agents in the clinic. In a second step, we point out how these changing modes of interaction also imply shifts in the conditions of trustworthiness, epistemic challenges regarding transparency, the underlying normative concepts of agency and its embedding into concrete contexts of deployment and, finally, the consequences for (possible) ascriptions of responsibility. Third, we draw first conclusions for further steps regarding a ‘meaningful human control’ of clinical AI-DSS. (shrink)

Decisions based on algorithmic, machine learning models can be unfair, reproducing biases in historical data used to train them. While computational techniques are emerging to address aspects of these concerns through communities such as discrimination-aware data mining and fairness, accountability and transparency machine learning, their practical implementation faces real-world challenges. For legal, institutional or commercial reasons, organisations might not hold the data on sensitive attributes such as gender, ethnicity, sexuality or disability needed to diagnose and mitigate emergent indirect discrimination-by-proxy, such (...) as redlining. Such organisations might also lack the knowledge and capacity to identify and manage fairness issues that are emergent properties of complex sociotechnical systems. This paper presents and discusses three potential approaches to deal with such knowledge and information deficits in the context of fairer machine learning. Trusted third parties could selectively store data necessary for performing discrimination discovery and incorporating fairness constraints into model-building in a privacy-preserving manner. Collaborative online platforms would allow diverse organisations to record, share and access contextual and experiential knowledge to promote fairness in machine learning systems. Finally, unsupervised learning and pedagogically interpretable algorithms might allow fairness hypotheses to be built for further selective testing and exploration. Real-world fairness challenges in machine learning are not abstract, constrained optimisation problems, but are institutionally and contextually grounded. Computational fairness tools are useful, but must be researched and developed in and with the messy contexts that will shape their deployment, rather than just for imagined situations. Not doing so risks real, near-term algorithmic harm. (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)

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)

This paper investigates the influence that social ties can have on behavior. After defining the concept of social ties that we consider, we introduce an original model of social ties. The impact of such ties on social preferences is studied in a coordination game with outside option. We provide a detailed game theoretical analysis of this game while considering various types of players, i.e., self-interest maximizing, inequity averse, and fair agents. In addition to these approaches that require strategic reasoning in (...) order to reach some equilibrium, we also present an alternative hypothesis that relies on the concept of team reasoning. After having discussed the differences between the latter and our model of social ties, we show how an experiment can be designed so as to discriminate among the models presented in the paper. (shrink)

Recent work on interpretability in machine learning and AI has focused on the building of simplified models that approximate the true criteria used to make decisions. These models are a useful pedagogical device for teaching trained professionals how to predict what decisions will be made by the complex system, and most importantly how the system might break. However, when considering any such model it’s important to remember Box’s maxim that "All models are wrong but some are useful." We focus on (...) the distinction between these models and explanations in philosophy and sociology. These models can be understood as a "do it yourself kit" for explanations, allowing a practitioner to directly answer "what if questions" or generate contrastive explanations without external assistance. Although a valuable ability, giving these models as explanations appears more difficult than necessary, and other forms of explanation may not have the same trade-offs. We contrast the different schools of thought on what makes an explanation, and suggest that machine learning might benefit from viewing the problem more broadly. (shrink)

The ethics of emerging forms of artificial intelligence has become a prolific subject in both academic and public spheres. A great deal of these concerns flow from the need to ensure that these technologies do not cause harm—physical, emotional or otherwise—to the human agents with which they will interact. In the literature, this challenge has been met with the creation of artificial moral agents: embodied or virtual forms of artificial intelligence whose decision procedures are constrained by explicit normative principles, requiring (...) the implementation of what is commonly called artificial morality into these agents. To date, the types of reasoning structures and principles which inform artificial morality have been of two kinds: first, an ethically maximal vision of artificial morality which relies on the strict implementation of traditional moral theories such as Kantian deontology or Utilitarianism, and second, a more minimalist vision which applies stochastic AI techniques to large data sets of human moral preferences so as to illicit or intuit general principles and preferences for the design of artificial morality. Taken individually, each approach is unable to fully answer the challenge of producing inoffensive behavior in artificial moral agents, most especially since both forms are unable to strike a balance between the ideal set of constraints which morality imposes on one hand, and the types of constraints public acceptability imposes, on the other. We provide an alternative approach to the design of artificial morality, the Ethical Valence Theory, whose purpose is to accommodate this balance, and apply this approach to the case of autonomous vehicles. (shrink)

In this panel, we explore the future of value sensitive design (VSD). The stakes are high. Many in public and private sectors and in civil society are gradually realizing that taking our values seriously implies that we have to ensure that values effectively inform the design of technology which, in turn, shapes people’s lives. Value sensitive design offers a highly developed set of theory, tools, and methods to systematically do so.

Ever since the use of artificial intelligence in the space sector, there have been progressive changes that brought both benefits and risks. AI technologies have started to influence human rights and freedoms, relationships with public authorities and the private sector. Therefore, the use of AI involves legal obligations caused by the influence and consequences of the dependency on the specified technologies. This indicates the need to investigate the issue of the legal problems and risks of using AI in space activity (...) to identify, describe and explain legal issues and dilemmas in general and in the space sector in particular. The author has formulated theoretical definitions in the defined area through the logical-semantic approach. The comparative method has allowed us to carry out a comparative analysis of the AI legal regulation in the space sector in the USA, China and the EU. The authors have gone into rethinking the issue of action, cause-effect relationship, legal personality in matters of prosecution for damage caused by autonomous space objects. As well as combining the existing space law and its state-oriented responsibility with the space activity of private players using autonomous objects. (shrink)

Conventional accounts of epistemic opacity, particularly those that stem from the definitive work of Paul Humphreys, typically point to limitations on the part of epistemic agents to account for the distinct ways in which systems, such as computational methods and devices, are opaque. They point, for example, to the lack of technical skill on the part of an agent, the failure to meet standards of best practice, or even the nature of an agent as reasons why epistemically relevant elements of (...) a process may be inaccessible. In this paper I argue that there are certain instances of epistemic opacity— particularly in computational methods such a computer simulations and machine learning processes—that do not arise from, are not responsive to, and are therefore not explained by the epistemic limitations of an agent. I call these instances agent-neutral and agent-independent instances of epistemic opacity respectively. As a result, I also argue that conventional accounts of epistemic opacity offer a limited understanding of the full spectrum of kinds and sources of epistemic opacity, particularly of the kind found in computational methods. In particular, as I will show below, the limitations of these accounts are reflected in the way they fail to provide satisfactory explanations when faced with certain instances of opacity. (shrink)