Responsibility Gap Research Articles

From Responsibility to Reason-Giving Explainable Artificial Intelligence

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.

Physical therapy within US HTCs: A multicentre survey of utilization, practice patterns and pain management approaches.

Practice patterns and utilization of physical therapists (PTs) affiliated with Hemophilia Treatment Centers (HTCs) in the United States (US) are not well known. Describe utilization, role responsibilities and practice patterns of US HTC PTs. Identify practice patterns specifically focusing on assessment and treatment of pain. Recognize gaps in utilization and role responsibilities of PTs as part of the multidisciplinary team and suggest guidelines for PT involvement within the HTC. Respondents were a subset of a convenience sample of healthcare providers who responded to a non-validated survey developed by a multi-disciplinary panel of haemophilia experts. A 33.0% response-rate (n=59) representing all regions of US HTCs was achieved. Those working ≥10hours per week were more likely to provide nutrition education (P=.026) and surgical options education (P<.001). Those who billed insurance for their services during comprehensive visits were more likely to provide education regarding surgical options (P=.046). The majority of PTs (95.0%) evaluated pain regardless of time spent in clinic and felt comfortable treating pain. Fifty-eight percent used a formal pain measurement tool and more likely to use a formal pain measurement tool if billing insurance (P=.004). Top five non-pharmacologic treatments recommended for pain management included splints/braces (84.8%), aquatic therapy (74.6%), orthotics (71.2%), surgical options (47.5%) and yoga (32.2%). This study demonstrated PT utilization across HTC centres varies widely. Gaps in care may be addressed through salary support, funded education, greater regional/national collaboration of PTs specializing in bleeding disorders and advocacy for insurance coverage for appropriate services.

Farm advisors amid the transition to Agriculture 4.0: Professional identity, conceptions of the future and future‐specific competencies

AbstractThe transition to Agriculture 4.0 creates new responsibilities for farm advisors and initiates changes to the professional trajectories. In this work, following a mixed research design, we examine how Greek advisors experience the transition to and anticipate the future of Agriculture 4.0. We also aim at identifying what elements they are changing in their professional identities to respond to the challenges associated with Agriculture 4.0, which are the new responsibility‐related competency needs that this transition generates, and how conceptions of the future of farming impact upon these needs. Our findings suggest that the transition to Agriculture 4.0 urges the emergence of a new culture, in which data and technology are considered more reliable than human advice, and creates new responsibility gaps. To cope with these changes, advisors add new roles in their professional identities, emphasising the principle of beneficence and paying limited attention to the societal externalities of transition. Advisors view Agriculture 4.0 as a threat or disruption more than as evolution or promise. In parallel, they prioritise different sets of competencies needed to responsibly facilitate agricultural digitalisation, depending on how they grasp the future of Agriculture 4.0.

Responsibility assignment won’t solve the moral issues of artificial intelligence

Who is responsible for the events and consequences caused by using artificially intelligent tools, and is there a gap between what human agents can be responsible for and what is being done using artificial intelligence? Both questions presuppose that the term ‘responsibility’ is a good tool for analysing the moral issues surrounding artificial intelligence. This article will draw this presupposition into doubt and show how reference to responsibility obscures the complexity of moral situations and moral agency, which can be analysed with a more differentiated toolset of moral terminology. It suggests that the impression of responsibility gaps only occurs if we gloss over the complexity of the moral situation in which artificial intelligent tools are employed and if—counterfactually—we ascribe them some kind of pseudo-agential status.

Diffused responsibility: attributions of responsibility in the use of AI-driven clinical decision support systems

Good decision-making is a complex endeavor, and particularly so in a health context. The possibilities for day-to-day clinical practice opened up by AI-driven clinical decision support systems (AI-CDSS) give rise to fundamental questions around responsibility. In causal, moral and legal terms the application of AI-CDSS is challenging existing attributions of responsibility. In this context, responsibility gaps are often identified as main problem. Mapping out the changing dynamics and levels of attributing responsibility, we argue in this article that the application of AI-CDSS causes diffusions of responsibility with respect to a causal, moral, and legal dimension. Responsibility diffusion describes the situation where multiple options and several agents can be considered for attributing responsibility. Using the example of an AI-driven ‘digital tumor board’, we illustrate how clinical decision-making is changed and diffusions of responsibility take place. Not denying or attempting to bridge responsibility gaps, we argue that dynamics and ambivalences are inherent in responsibility, which is based on normative considerations such as avoiding experiences of disregard and vulnerability of human life, which are inherently accompanied by a moment of uncertainty, and is characterized by revision openness. Against this background and to avoid responsibility gaps, the article concludes with suggestions for managing responsibility diffusions in clinical decision-making with AI-CDSS.

Applying a Common Enterprise Theory of Liability to Clinical AI Systems.

The advent of artificial intelligence ("AI") holds great potential to improve clinical diagnostics. At the same time, there are important questions of liability for harms arising from the use of this technology. Due to their complexity, opacity, and lack of foreseeability, AI systems are not easily accommodated by traditional liability frameworks. This difficulty is compounded in the health care space where various actors, namely physicians and health care organizations, are subject to distinct but interrelated legal duties regarding the use of health technology. Without a principled way to apportion responsibility among these actors, patients may find it difficult to recover for injuries. In this Article, I propose that physicians, manufacturers of clinical AI systems, and hospitals be considered a common enterprise for the purposes of liability. This proposed framework helps facilitate the apportioning of responsibility among disparate actors under a single legal theory. Such an approach responds to concerns about the responsibility gap engendered by clinical AI technology as it shifts away from individualistic notions of responsibility, embodied by negligence and products liability, toward a more distributed conception. In addition to favoring plaintiff recovery, a common enterprise strict liability approach would create strong incentives for the relevant actors to take care.

Mind the gap: independent directors and corporate social responsibility in China

PurposeThis study aims to examine the effect of independent directors on the corporate social responsibility (CSR) gap – a misalignment between internal and external CSR. More specifically, the authors investigate how two types of independent directors (i.e. politically connected and foreign) affect a firm’s CSR gap in China.Design/methodology/approachThe authors use the fixed-effects regression model to analyze the panel dataset, which is conducted by a sample of Chinese publicly listed firms from 2008 to 2015.FindingsThe findings indicate that, on average, firms undertake more external than internal CSR actions. Importantly, the authors find that firms having politically connected independent directors on boards have a wider gap between their internal and external CSR. In contrast, firms having foreign independent directors on boards have a narrower gap between their internal and external CSR.Practical implicationsThis study provides insights into the role of independent directors in increasing or decreasing the gap between a firm’s internal and external CSR actions, which offers important implications for policymakers and investors.Originality/valueThis study extends the literature on the causes of the CSR gap and deepens the theoretical understanding of the governance role of independent directors in China.

The Accountability of Software Developers for War Crimes Involving Autonomous Weapons

This Article considers the extent to which the joint criminal enterprise doctrine could be invoked to hold software developers criminally accountable for violations of international humanitarian law involving autonomous weapons. More specifically, it considers whether the third part of the concept—which concerns common criminal purposes—might be brought to bear to achieve this end. The doctrine is deconstructed into five components, and each component is analyzed both in abstract and in terms of practical application. The Article establishes that, in certain contexts, software developers can and should be held accountable through this mechanism. Thus, it demonstrates that it is possible to avoid the emergence of a “responsibility gap” if, or more likely when, autonomous weapons with offensive capabilities are finally deployed on the battlefield.&nbsp;

Understanding Responsibility in Stem Cell Research

"Over the years, numerous ethical implications in stem cell research have been identified. Consequentially, there is a need to anticipate, prevent and/or mitigate these implications. In literature and in the guidelines of the International Society for Stem Cell Research some of these implications have been reframed into (moral) responsibilities. What exactly is meant by responsibility and which notions of responsibility are important often remains unclear. As a consequence, this encumbers the possibility to deal with these responsibilities in a systematic way. Therefore, the goal of this paper is to discuss how the concept of responsibility in stem cell research could best be understood. This paper addresses which notions of responsibility are relevant for the field of stem cell research. This will be done by first distinguishing between backward-looking and forward-looking notions of responsibilities, where backward-looking responsibilities are about reactive attitudes and forward-looking responsibilities are about what could be expected from someone to prevent ethical implications. Subsequently, ethical implications of and wrongdoings in stem cell research will be categorized in these notions of responsibility. Taking one step back and looking at the different notions of responsibility could help to identify gaps in responsibilities as well as to distinguish obligatory and supererogatory responsibilities. Consequentially, this informs how to prioritize, distribute and delegate responsibilities over the different stakeholders in stem cell research. This paper concludes by discussing the distribution of responsibilities and different strategies to promote responsibility in stem cell research. "

O-098 Embryo selection using Artificial Intelligence (AI): Epistemic and ethical considerations

Abstract Study question What are the epistemic and ethical considerations of clinically implementing Artificial Intelligence (AI) algorithms in embryo selection? Summary answer AI embryo selection algorithms used to date are “black-box” models with significant epistemic and ethical issues, and there are no trials assessing their clinical effectiveness. What is known already The innovation of time-lapse imaging offers the potential to generate vast quantities of data for embryo assessment. Computer Vision allows image data to be analysed using algorithms developed via machine learning which learn and adapt as they are exposed to more data. Most algorithms are developed using neural networks and are uninterpretable (or “black box”). Uninterpretable models are either too complicated to understand or proprietary, in which case comprehension is impossible for outsiders. In the IVF context, these outsiders include doctors, embryologists and patients, which raises ethical questions for its use in embryo selection. Study design, size, duration We performed a scoping review of articles evaluating AI for embryo selection in IVF. We considered the epistemic and ethical implications of current approaches. Participants/materials, setting, methods We searched Medline, Embase, ClinicalTrials.gov and the EU Clinical Trials Register for full text papers evaluating AI for embryo selection using the following key words: artificial intelligence* OR AI OR neural network* OR machine learning OR support vector machine OR automatic classification AND IVF OR in vitro fertilisation OR embryo*, as well as relevant MeSH and Emtree terms for Medline and Embase respectively. Main results and the role of chance We found no trials evaluating clinical effectiveness either published or registered. We found efficacy studies which looked at 2 types of outcomes – accuracy for predicting pregnancy or live birth and agreement with embryologist evaluation. Some algorithms were shown to broadly differentiate well between “good-” and “poor-” quality embryos but not between embryos of similar quality, which is the clinical need. Almost universally, the AI models were opaque (“black box”) in that at least some part of the process was uninterpretable. “Black box” models are problematic for epistemic and ethical reasons. Epistemic concerns include information asymmetries between algorithm developers and doctors, embryologists and patients; the risk of biased prediction caused by known and/or unknown confounders during the training process; difficulties in real-time error checking due to limited interpretability; the economics of buying into commercial proprietary models, brittle to variation in the treatment process; and an overall difficulty troubleshooting. Ethical pitfalls include the risk of misrepresenting patient values; concern for the health and well-being of future children; the risk of disvaluing disability; possible societal implications; and a responsibility gap, in the event of adverse events. Limitations, reasons for caution Our search was limited to the two main medical research databases. Although we checked article references for more publications, we were less likely to identify studies that were not indexed in Medline or Embase, especially if they were not cited in studies identified in our search. Wider implications of the findings It is premature to implement AI for embryo selection outside of a clinical trial. AI for embryo selection is potentially useful, but must be done carefully and transparently, as the epistemic and ethical issues are significant. We advocate for the use of interpretable AI models to overcome these issues. Trial registration number not applicable

Liability Attribution in the Context of AI Use

AI (artificial intelligence) systems represent a double-bladed sword. AI brings both benefits and harms to our society. When adverse consequences are produced by AI use in practice, responsibility gaps emerge, due to the opacity and complex nature of these sociotechnical systems. In this study, we propose that assigning liabilities collectively among AI actors can bridge the responsibility gap, hence promoting justice and ethical uses of AI in society. We do so by drawing on the field of philosophy and attempting to answer the question of how (to whom and to what extent) responsibility can be allocated to reduce accountability gaps in the AI context. We generate a set of propositions that lay out the ground for future research this important interdisciplinary domain.

STI policy and governance in Sub-Saharan Africa: Fostering actors’ interactions in research and innovation

This study investigates science, technology, and innovation (STI) policy and governance in relation to research and innovation (R&amp;I) ecosystems in Sub-Saharan Africa (SSA). To understand the implications of STI policy and governance on R&amp;I, the authors focus on university, industry and government actors; using the Triple Helix and National Systems of Innovation approaches as analytical frameworks to guide the study. The authors identify a range of factors hindering R&amp;I actors’ interactions, including gaps in STI governance responsibilities and accountabilities, policymaking and structural transitions, R&amp;I management capacity and capability gaps. Based on the findings, they propose changes to policymaking and the governance of R&amp;I in SSA, in addition to advancing innovative approaches such as ‘resource circulation’ in the context of knowledge, research and science infrastructure scarcity. The findings provide fresh insights into STI policy and governance frameworks, fostering actors’ interactions and supporting performance improvements in research, science and innovation systems across Africa.

Artificial Moral Responsibility: How We Can and Cannot Hold Machines Responsible.

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 widening responsibility 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 of agency or 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 of artificial moral responsibility.

Four Responsibility Gaps with Artificial Intelligence: Why they Matter and How to Address them

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

Autonomous Weapons and the Responsibility Gap in light of the Mens Rea of the War Crime of Attacking Civilians in the ICC Statute

Abstract Within the broader context of the problems raised by the interaction between humans and machines in weaponry and targeting, this paper deals with the specific issue of the mens rea required to establish responsibility for the war crime of indiscriminate attacks, in the context of attacks performed with semi-autonomous weapons or with the support of artificial intelligence (AI) in targeting decision-making. The author presents the difficulties that are determined by the interaction between humans and machines, and highlights that an interpretation that would allow for risk-taking mental elements such as dolus eventualis and recklessness in the framework of the war crime of attacking civilians would be better able to capture the criminality of the conduct of the person who knowingly accepts the risk of killing civilians as part of an AI-powered attack where the result of hitting the civilian target is one of the possible outcomes. However, the article indicates that this construction can be employed only in specific circumstances, since in most scenarios even these lowered mens rea requirements would not be met. In most human-machine teaming scenarios, lower types of intent such as dolus eventualis would still be insufficient for the ascription of criminal responsibility for such indiscriminate attacks against civilians. This is because of the specific risks posed by the integration of autonomy in the targeting process and the resulting changes to the cognitive environment in which human agents operate, which significantly affect specific components of mens rea standards.

The Spectrum of Responsibility Ascription for End Users of Neurotechnologies

Invasive neural devices offer novel prospects for motor rehabilitation on different levels of agentive behavior. From a functional perspective, they interact with, support, or enable human intentional actions in such a way that movement capabilities are regained. However, when there is a technical malfunction resulting in an unintended movement, the complexity of the relationship between the end user and the device sometimes makes it difficult to determine who is responsible for the outcome - a circumstance that has been coined as "responsibility gap" in the literature. So far, recent accounts frame this issue around the theme of control but more work is needed to explore the complicated terrain of assigning responsibility for neural device-mediated actions from this control perspective. This paper aims at contributing to this tendency by offering more fine-grained distinctions of how that control capacity is facilitated by the machine and how it can be exercised by the end user. This results in a novel framework that depicts an in-depth exploration of the control aspect of responsibility in a way that incorporates the diversity of relationships between neurotechnologies, the various conditions they treat, and the individual end user's experience.

Gaps in Hospital and Skilled Nursing Facility Responsibilities During Transitions of Care: a Comparison of Hospital and SNF Clinicians' Perspectives.

Adverse outcomes are common in transitions from hospital to skilled nursing facilities (SNFs). Gaps in transitional care processes contribute to these outcomes, but it is unclear whether hospital and SNF clinicians have the same perception about who is responsible for filling these gaps in care transitions. We sought to understand the perspectives of hospital and SNF clinicians on their roles and responsibilities in transitional care processes, to identify areas of congruence and gaps that could be addressed to improve transitions. Semi-structured interviews with interdisciplinary hospital and SNF providers. Forty-one clinicians across 3 hospitals and 3 SNFs including nurses (8), social workers (7), physicians (8), physical and occupational therapists (12), and other staff (6). Using team-based approach to deductive analysis, we mapped responses to the 10 domains of the Ideal Transitions of Care Framework (ITCF) to identify areas of agreement and gaps between hospitals and SNFs. Although both clinician groups had similar conceptions of an ideal transitions of care, their perspectives included significant gaps in responsibilities in 8 of the 10 domains of ITCF, including Discharge Planning; Complete Communication of Information; Availability, Timeliness, Clarity and Organization of Information; Medication Safety; Educating Patients to Promote Self-Management; Enlisting Help of Social and Community Supports; Coordinating Care Among Team Members; and Managing Symptoms After Discharge. As hospitals and SNFs increasingly are held jointly responsible for the outcomes of patients transitioning between them, clarity in roles and responsibilities between hospital and SNF staff are needed. Improving transitions of care may require site-level efforts, joint hospital-SNF initiatives, and national financial, regulatory, and technological fixes. In the meantime, building effective hospital-SNF partnerships is increasingly important to delivering high-quality care to a vulnerable older adult population.

Moral responsibility for computationally designed products

Computational design systems (such as those using evolutionary algorithms) can create designs for a variety of physical products. Introducing these systems into the design process risks creating a ‘responsibility gap’ for flaws in the products they are used to create, as human designers may no longer believe that they are wholly responsible for them. We respond to this problem by distinguishing between causal responsibility and capacity responsibility (the ability to be morally responsible for actions) for creating product designs to argue that while the computational design systems and human designers are both casually responsible for creating product designs, the human designers who use these systems and the developers who create them have capacity responsibility for such designs. We show that there is no responsibility gap for products designed using computational design systems by comparing different accounts of moral responsibility for robots and AI (instrumentalism, machine ethics, and hybrid responsibility). We argue that all three of these accounts of moral responsibility for AI systems support the conclusion that the product designers who use computational design systems and the developers of these systems are morally responsible for any flaws or faults in the products designed by these systems. We conclude by showing how the responsibilities of accountability and blameworthiness should be attributed between the product designers, the developers of the computational design systems.

Artificial intelligence and responsibility

In the debate on whether to ban LAWS, moral arguments are mainly used. One of these arguments, proposed by Sparrow, is that the use of LAWS goes hand in hand with the responsibility gap. Together with the premise that the ability to hold someone responsible is a necessary condition for the admissibility of an act, Sparrow believes that this leads to the conclusion that LAWS should be prohibited. In this article, it will be shown that Sparrow’s argumentation for both premises is not convincing. If one interprets the thesis that responsibility (first premise) is necessary in a descriptive sense, this assertion clashes with military theory and practice. And even if you focus on the normative interpretation, that claim does not stand. The second premise for Sparrow’s conclusion, namely that you cannot hold anyone responsible for LAWS’ (mis)deeds, is based on the idea that control is a necessary condition for responsibility. It will be shown that this idea too is not correct, which means that Sparrow’s control argument does not do the work it should do. From this, we can conclude that Sparrow’s justification for his claim that LAWS should be banned is insufficient, and neither can we conclude that the thesis of a responsibility gap has in any case been undermined. However, it will also be argued that someone may be responsible for the actions of LAWS, or that it cannot be excluded that one can be held responsible.

From Automation to Autonomy: Legal and Ethical Responsibility Gaps in Artificial Intelligence Innovation

The increasing prominence of artificial intelligence (AI) systems in daily life and the evolving capacity of these systems to process data and act without human input raise important legal and ethical concerns. This article identifies three primary AI actors in the value chain (innovators, providers, and users) and three primary types of AI (automation, augmentation, and autonomy). It then considers responsibility in AI innovation from two perspectives: (i) strict liability claims arising out of the development, commercialization, and use of products with built-in AI capabilities (designated herein as “AI artifacts”); and (ii) an original research study on the ethical practices of developers and managers creating AI systems and AI artifacts. The ethical perspective is important because, at the moment, the law is poised to fall behind technological reality—if it hasn’t already. Consideration of the liability issues in tandem with ethical perspectives yields a more nuanced assessment of the likely consequences and adverse impacts of AI innovation. Companies should consider both legal and ethical strategies thinking about their own liability and ways to limit it, as well as policymakers considering AI regulation ex ante.