Healthcare (including public health) is special because it protects normal functioning, which in turn protects the range of opportunities open to individuals. I extend this account in two ways. First, since the distribution of goods other than healthcare affect population health and its distribution, I claim that Rawls's principles of justice describe a fair distribution of the social determinants of health, giving a partial account of when health inequalities are unjust. Second, I supplement a principled account of justice for health (...) and healthcare with an account of fair process for setting limits or rationing care. This account is provided by three conditions that comprise "accountability for reasonableness.". (shrink)

Scientists depend on complex computational systems that are often ineliminably opaque, to the detriment of our ability to give scientific explanations and detect artifacts. Some philosophers have s...

This article examines the complaint that arbitrary algorithmic decisions wrong those whom they affect. It makes three contributions. First, it provides an analysis of what arbitrariness means in this context. Second, it argues that arbitrariness is not of moral concern except when special circumstances apply. However, when the same algorithm or different algorithms based on the same data are used in multiple contexts, a person may be arbitrarily excluded from a broad range of opportunities. The third contribution is to explain (...) why this systemic exclusion is of moral concern and to offer a solution to address it. (shrink)

How should one react when one has a belief, but knows that other people—who have roughly the same evidence as one has, and seem roughly as likely to react to it correctly—disagree? This paper argues that the disagreement of other competent inquirers often requires one to be much less confident in one’s opinions than one would otherwise be.

A number of findings in the field of machine learning have given rise to questions about what it means for automated scoring- or decisionmaking systems to be fair. One center of gravity in this discussion is whether such systems ought to satisfy classification parity (which requires parity in accuracy across groups, defined by protected attributes) or calibration (which requires similar predictions to have similar meanings across groups, defined by protected attributes). Central to this discussion are impossibility results, owed to Kleinberg (...) et al. (2016), Chouldechova (2017), and Corbett-Davies et al. (2017), which show that classification parity and calibration are often incompatible. This paper aims to argue that classification parity, calibration, and a newer, interesting measure called counterfactual fairness are unsatisfactory measures of fairness, offer a general diagnosis of the failure of these measures, and sketch an alternative approach to understanding fairness in machine learning. (shrink)

In lieu of an abstract, here is a brief excerpt of the content:The Ideal of Shared Decision Making Between Physicians and PatientsDan W. Brock (bio)IntroductionShared treatment decision making, with its division of labor between physician and patient, is a common ideal in medical ethics for the physician-patient relationship.1 Most simply put, the physician's role is to use his or her training, knowledge, and experience to provide the patient with facts about the diagnosis and about the prognoses without treatment and with (...) alternative treatments. The patient's role in this division of labor is to provide the values—his or her own conception of the good—with which to evaluate these alternatives, and to select the one that is best for himself or herself. As a rough guide to practice, this is a reasonable conception; most of the time it is likely to produce sound treatment decisions. However, as an ideal it is too simplistic, and is subject to several challenges that I will explore in this paper.Some challenges relate to the physician's role. This facts/values division of labor seems to assume that the physician can and should provide the facts about treatment alternatives in a value-neutral form. But some have questioned whether the sciences on which medicine is based are, or can be, value free. Moreover, the concepts of health and disease, and of the normal and pathological, are held by many to be value laden. And even if it is possible for physicians to provide only value-neutral facts to the patient, should the physicians' role be restricted in this way?Other challenges that I will present in the paper are to the patient's role [End Page 28] as provider of the values for the evaluation of the different treatment alternatives. One defense of this role is based on the claim that physicians are not in a position to know reliably what is in a patient's best interests (Buchanan, forthcoming). I believe at least as influential—even if often not explicitly stated—is an extreme subjectivism about values that assumes that the patient's own ultimate values, which define his or her own conception of the good, are incorrigible. By incorrigible, I mean they cannot be mistaken. In the standard case in which treatment is pursued for the patient's benefit, this extreme subjectivism is thought to support the patient's values guiding treatment decision making and treatment. I will explore this incorrigibility claim and will argue that on each of the main philosophical conceptions of the good for persons it is indefensible and should not be the basis for shared treatment decision making.Underlying this view of the physician-patient division of labor are assumptions and beliefs about the nature and relation of facts and values, and science and ethics, that are largely the legacy of a logical positivism that has long since been rejected by most philosophers. It is time that these unstable foundations for our normative ideal of the physician-patient relation were finally removed and replaced with more defensible underpinnings.Two Models of the Physician-Patient RelationshipPositivists insisted on a relatively sharp distinction between descriptive or empirical claims and science, on the one hand, and evaluative claims and ethics, on the other.2 Descriptive statements and the claims of science were to be true or false according to whether they in fact correctly described and explained the world. By remaining properly descriptive or empirical, science could, at least in principle, be entirely value neutral or value free.In its more extreme versions, positivism held that ethical judgments lacked cognitive content, but were instead expressions of emotions or attitudes, and as such could be neither true nor false, correct or mistaken. Because no evaluative statements were held to be logically entailed by any descriptive or empirical statements, moral reasoning was thought to be properly understood not as reasoning but as attempts at nonrational persuasion.3 Many drew from this an extreme subjectivism in ethics along the lines of what I have called above the incorrigibility thesis—there is no meaningful sense in which an individual's ultimate values defining his or her own conception of the good life could be mistaken, false, or unfounded.These... (shrink)

This paper argues that data-driven medicine gives rise to a particular normative challenge. Against the backdrop of a distinction between the good and the right, harnessing personal health data towards the development and refinement of data-driven medicine is to be welcomed from the perspective of the good. Enacting solidarity drives progress in research and clinical practice. At the same time, such acts of sharing could—especially considering current developments in big data and artificial intelligence—compromise the right by leading to injustices and (...) affecting concrete modes of individual self-determination. In order to address this potential tension, two key elements for ethical reflection on data-driven medicine are proposed: the controllability of information flows, including technical infrastructures that are conducive towards controllability, and a paradigm shift towards output-orientation in governance and policy. (shrink)

Deep neural networks have become increasingly successful in applications from biology to cosmology to social science. Trained DNNs, moreover, correspond to models that ideally allow the prediction of new phenomena. Building in part on the literature on ‘eXplainable AI’, I here argue that these models are instrumental in a sense that makes them non-explanatory, and that their automated generation is opaque in a unique way. This combination implies the possibility of an unprecedented gap between discovery and explanation: When unsupervised models (...) are successfully used in exploratory contexts, scientists face a whole new challenge in forming the concepts required for understanding underlying mechanisms. (shrink)

Advanced AI systems are rapidly making their way into medical research and practice, and, arguably, it is only a matter of time before they will surpass human practitioners in terms of accuracy, reliability, and knowledge. If this is true, practitioners will have a prima facie epistemic and professional obligation to align their medical verdicts with those of advanced AI systems. However, in light of their complexity, these AI systems will often function as black boxes: the details of their contents, calculations, (...) and procedures cannot be meaningfully understood by human practitioners. When AI systems reach this level of complexity, we can also speak of black-box medicine. In this paper, we want to argue that black-box medicine conflicts with core ideals of patient-centered medicine. In particular, we claim, black-box medicine is not conducive for supporting informed decision-making based on shared information, shared deliberation, and shared mind between practitioner and patient. (shrink)

Recent scholarship in philosophy of science and technology has shown that scientific and technological decision making are laden with values, including values of a social, political, and/or ethical character. This paper examines the role of value judgments in the design of machine-learning systems generally and in recidivism-prediction algorithms specifically. Drawing on work on inductive and epistemic risk, the paper argues that ML systems are value laden in ways similar to human decision making, because the development and design of ML systems (...) requires human decisions that involve tradeoffs that reflect values. In many cases, these decisions have significant—and, in some cases, disparate—downstream impacts on human lives. After examining an influential court decision regarding the use of proprietary recidivism-prediction algorithms in criminal sentencing, Wisconsin v. Loomis, the paper provides three recommendations for the use of ML in penal systems. (shrink)

Bioethics, Volume 36, Issue 2, Page 121-133, February 2022.

Should decision-making algorithms be held to higher standards of transparency than human beings? The way we answer this question directly impacts what we demand from explainable algorithms, how we govern them via regulatory proposals, and how explainable algorithms may help resolve the social problems associated with decision making supported by artificial intelligence. Some argue that algorithms and humans should be held to the same standards of transparency and that a double standard of transparency is hardly justified. We give two arguments (...) to the contrary and specify two kinds of situations for which higher standards of transparency are required from algorithmic decisions as compared to humans. Our arguments have direct implications on the demands from explainable algorithms in decision-making contexts such as automated transportation. (shrink)

One of the main difficulties in assessing artificial intelligence (AI) is the tendency for people to anthropomorphise it. This becomes particularly problematic when we attach human moral activities to AI. For example, the European Commission’s High-level Expert Group on AI (HLEG) have adopted the position that we should establish a relationship of trust with AI and should cultivate trustworthy AI (HLEG AI Ethics guidelines for trustworthy AI, 2019, p. 35). Trust is one of the most important and defining activities in (...) human relationships, so proposing that AI should be trusted, is a very serious claim. This paper will show that AI cannot be something that has the capacity to be trusted according to the most prevalent definitions of trust because it does not possess emotive states or can be held responsible for their actions—requirements of the affective and normative accounts of trust. While AI meets all of the requirements of the rational account of trust, it will be shown that this is not actually a type of trust at all, but is instead, a form of reliance. Ultimately, even complex machines such as AI should not be viewed as trustworthy as this undermines the value of interpersonal trust, anthropomorphises AI, and diverts responsibility from those developing and using them. (shrink)

We are sceptical of concerns over the opacity of algorithmic decision tools. While transparency and explainability are certainly important desiderata in algorithmic governance, we worry that automated decision-making is being held to an unrealistically high standard, possibly owing to an unrealistically high estimate of the degree of transparency attainable from human decision-makers. In this paper, we review evidence demonstrating that much human decision-making is fraught with transparency problems, show in what respects AI fares little worse or better and argue that (...) at least some regulatory proposals for explainable AI could end up setting the bar higher than is necessary or indeed helpful. The demands of practical reason require the justification of action to be pitched at the level of practical reason. Decision tools that support or supplant practical reasoning should not be expected to aim higher than this. We cast this desideratum in terms of Daniel Dennett’s theory of the “intentional stance” and argue that since the justification of action for human purposes takes the form of intentional stance explanation, the justification of algorithmic decisions should take the same form. In practice, this means that the sorts of explanations for algorithmic decisions that are analogous to intentional stance explanations should be preferred over ones that aim at the architectural innards of a decision tool. (shrink)

Journal of Political Philosophy, Volume 30, Issue 2, Page 209-229, June 2022.

Technologies that deploy data science methods are liable to result in epistemic harms involving the diminution of individuals with respect to their standing as knowers or their credibility as sources of testimony. Not all harms of this kind are unjust but when they are we ought to try to prevent or correct them. Epistemically unjust harms will typically intersect with other more familiar and well-studied kinds of harm that result from the design, development, and use of data science technologies. However, (...) we argue that epistemic injustices can be distinguished conceptually from more familiar kinds of harm. We argue that epistemic harms are morally relevant even in cases where those who suffer them are unharmed in other ways. Via a series of examples from the criminal justice system, workplace hierarchies, and educational contexts we explain the kinds of epistemic injustice that can result from common uses of data science technologies. (shrink)

Simple idealized models seem to provide more understanding than opaque, complex, and hyper-realistic models. However, an increasing number of scientists are going in the opposite direction by utilizing opaque machine learning models to make predictions and draw inferences, suggesting that scientists are opting for models that have less potential for understanding. Are scientists trading understanding for some other epistemic or pragmatic good when they choose a machine learning model? Or are the assumptions behind why minimal models provide understanding misguided? In (...) this paper, using the case of deep neural networks, I argue that it is not the complexity or black box nature of a model that limits how much understanding the model provides. Instead, it is a lack of scientific and empirical evidence supporting the link that connects a model to the target phenomenon that primarily prohibits understanding. (shrink)

Machine Learning (ML) is on the rise in medicine, promising improved diagnostic, therapeutic and prognostic clinical tools. While these technological innovations are bound to transform health care, they also bring new ethical concerns to the forefront. One particularly elusive challenge regards discriminatory algorithmic judgements based on biases inherent in the training data. A common line of reasoning distinguishes between justified differential treatments that mirror true disparities between socially salient groups, and unjustified biases which do not, leading to misdiagnosis and erroneous (...) treatment. In the curation of training data this strategy runs into severe problems though, since distinguishing between the two can be next to impossible. We thus plead for a pragmatist dealing with algorithmic bias in healthcare environments. By recurring to a recent reformulation of William James’s pragmatist understanding of truth, we recommend that, instead of aiming at a supposedly objective truth, outcome-based therapeutic usefulness should serve as the guiding principle for assessing ML applications in medicine. (shrink)

This paper develops an account of trustworthy AI. Its central idea is that whether AIs are trustworthy is a matter of whether they live up to their function-based obligations. We argue that this account serves to advance the literature in a couple of important ways. First, it serves to provide a rationale for why a range of properties that are widely assumed in the scientific literature, as well as in policy, to be required of trustworthy AI, such as safety, justice, (...) and explainability, are properties (often) instantiated by trustworthy AI. Second, we connect the discussion on trustworthy AI in policy, industry, and the sciences with the philosophical discussion of trustworthiness. We argue that extant accounts of trustworthiness in the philosophy literature cannot make proper sense of trustworthy AI and that our account compares favourably with its competitors on this front. (shrink)

The datafication and digitalization of health and medicine has engendered a proliferation of new collaborations between public health institutions and data corporations like Google, Apple, Microsoft and Amazon. Critical perspectives on these new partnerships tend to frame them as an instance of market transgressions by tech giants into the sphere of health and medicine, in line with a “hostile worlds” doctrine that upholds that the borders between market and non-market spheres should be carefully policed. This article seeks to outline the (...) limitations of this common framing for critically understanding the phenomenon of the Googlization of health. In particular, the mobilization of a diversity of non-market value statements in the justification work carried out by actors involved in the Googlization of health indicates the co-presence of additional worlds or spheres in this context, which are not captured by the market vs. non-market dichotomy. It then advances an alternative framework, based on a multiple-sphere ontology that draws on Boltanski and Thevenot’s orders of worth and Michael Walzer’s theory of justice, which I call a normative pragmatics of justice. This framework addresses both the normative deficit in Boltanski and Thevenot’s work and provides an important emphasis on the empirical workings of justice. Finally, I discuss why this framework is better equipped to identify and to address the many risks raised by the Googlization of health and possibly other dimensions of the digitalization and datafication of society. (shrink)

We argue that the health sciences make causal claims on the basis of evidence both of physical mechanisms, and of probabilistic dependencies. Consequently, an analysis of causality solely in terms of physical mechanisms or solely in terms of probabilistic relationships, does not do justice to the causal claims of these sciences. Yet there seems to be a single relation of cause in these sciences - pluralism about causality will not do either. Instead, we maintain, the health sciences require a theory (...) of causality that unifies its mechanistic and probabilistic aspects. We argue that the epistemic theory of causality provides the required unification. (shrink)

Explainable artificial intelligence (XAI) is an emerging, multidisciplinary field of research that seeks to develop methods and tools for making AI systems more explainable or interpretable. XAI researchers increasingly recognise explainability as a context-, audience- and purpose-sensitive phenomenon, rather than a single well-defined property that can be directly measured and optimised. However, since there is currently no overarching definition of explainability, this poses a risk of miscommunication between the many different researchers within this multidisciplinary space. This is the problem we (...) seek to address in this paper. We outline a framework, called Explanatory Pragmatism, which we argue has two attractive features. First, it allows us to conceptualise explainability in explicitly context-, audience- and purpose-relative terms, while retaining a unified underlying definition of explainability. Second, it makes visible any normative disagreements that may underpin conflicting claims about explainability regarding the purposes for which explanations are sought. Third, it allows us to distinguish several dimensions of AI explainability. We illustrate this framework by applying it to a case study involving a machine learning model for predicting whether patients suffering disorders of consciousness were likely to recover consciousness. (shrink)

Machine learning (ML) systems play an increasingly relevant role in medicine and healthcare. As their applications move ever closer to patient care and cure in clinical settings, ethical concerns about the responsibility of their use come to the fore. I analyse an aspect of responsible ML use that bears not only an ethical but also a significant epistemic dimension. I focus on ML systems’ role in mediating patient–physician relations. I thereby consider how ML systems may silence patients’ voices and relativise (...) the credibility of their opinions, which undermines their overall credibility status without valid moral and epistemic justification. More specifically, I argue that withholding credibilitydue tohow ML systems operate can be particularly harmful to patients and, apart from adverse outcomes, qualifies as a form of testimonial injustice. I make my case for testimonial injustice in medical ML by considering ML systems currently used in the USA to predict patients’ risk of misusing opioids (automated Prediction Drug Monitoring Programmes, PDMPs for short). I argue that the locus of testimonial injustice in ML-mediated medical encounters is found in the fact that these systems are treated asmarkers of trustworthinesson which patients’ credibility is assessed. I further show how ML-based PDMPs exacerbate and further propagate social inequalities at the expense of vulnerable social groups. (shrink)

Simion and Kelp defend a non-anthropocentric account of trustworthy AI, based on the idea that the obligations of AI systems should be sourced in purely functional norms. In this commentary, I highlight some pressing counterexamples to their account, involving AI systems that reliably fulfil their functions but are untrustworthy because those functions are antagonistic to the interests of the trustor. Instead, I outline an alternative account, based on the idea that AI systems should not be considered primarily as tools but (...) as technological participants in social practices. Specifically, I propose to source the obligations of an AI system in the norms that should govern the role it plays within the social practices it participates in, taking into account any changes to the social practices that its participation may bring about. This proposal is anthropocentric insofar as it ascribes obligations to AI systems that are similar to those of human participants in social practices, but only modestly so, as it does not require trustworthy AI to have contentious anthropomorphic capacities (e.g. for consciousness or moral responsibility). (shrink)

Self‐driving cars hold out the promise of being much safer than regular cars. Yet they cannot be 100% safe. Accordingly, we need to think about who should be held responsible when self‐driving cars crash and people are injured or killed. We also need to examine what new ethical obligations might be created for car users by the safety potential of self‐driving cars. The article first considers what lessons might be learned from the growing legal literature on responsibility for crashes with (...) self‐driving cars. Next, worries about responsibility gaps and retribution gaps from the philosophical literature are introduced. This leads to a discussion of whether self‐driving cars are a form of agents that act independently of human agents. It is suggested that it is better to analyze their apparent agency in terms of human–robot collaborations, within which humans play the most important roles. The next topic is the idea that the safety potential of self‐driving cars might create a duty to either switch to self‐driving cars or seek means of making conventional cars safer. Lastly, there is a short discussion of ethical issues related to safe human–robot coordination within mixed traffic featuring both self‐driving cars and conventional cars. (shrink)

This paper explores the important role of critical science, and in particular of post-colonial and decolonial theories, in understanding and shaping the ongoing advances in artificial intelligence. Artificial intelligence is viewed as amongst the technological advances that will reshape modern societies and their relations. While the design and deployment of systems that continually adapt holds the promise of far-reaching positive change, they simultaneously pose significant risks, especially to already vulnerable peoples. Values and power are central to this discussion. Decolonial theories (...) use historical hindsight to explain patterns of power that shape our intellectual, political, economic, and social world. By embedding a decolonial critical approach within its technical practice, AI communities can develop foresight and tactics that can better align research and technology development with established ethical principles, centring vulnerable peoples who continue to bear the brunt of negative impacts of innovation and scientific progress. We highlight problematic applications that are instances of coloniality, and using a decolonial lens, submit three tactics that can form a decolonial field of artificial intelligence: creating a critical technical practice of AI, seeking reverse tutelage and reverse pedagogies, and the renewal of affective and political communities. The years ahead will usher in a wave of new scientific breakthroughs and technologies driven by AI research, making it incumbent upon AI communities to strengthen the social contract through ethical foresight and the multiplicity of intellectual perspectives available to us, ultimately supporting future technologies that enable greater well-being, with the goal of beneficence and justice for all. (shrink)

Artificial intelligence is increasingly being developed for use in medicine, including for diagnosis and in treatment decision making. The use of AI in medical treatment raises many ethical issues that are yet to be explored in depth by bioethicists. In this paper, I focus specifically on the relationship between the ethical ideal of shared decision making and AI systems that generate treatment recommendations, using the example of IBM’s Watson for Oncology. I argue that use of this type of system creates (...) both important risks and significant opportunities for promoting shared decision making. If value judgements are fixed and covert in AI systems, then we risk a shift back to more paternalistic medical care. However, if designed and used in an ethically informed way, AI could offer a potentially powerful way of supporting shared decision making. It could be used to incorporate explicit value reflection, promoting patient autonomy. In the context of medical treatment, we need value-flexible AI that can both respond to the values and treatment goals of individual patients and support clinicians to engage in shared decision making. (shrink)

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

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

Although decision‐making algorithms are not new to medicine, the availability of vast stores of medical data, gains in computing power, and breakthroughs in machine learning are accelerating the pace of their development, expanding the range of questions they can address, and increasing their predictive power. In many cases, however, the most powerful machine learning techniques purchase diagnostic or predictive accuracy at the expense of our ability to access “the knowledge within the machine.” Without an explanation in terms of reasons or (...) a rationale for particular decisions in individual cases, some commentators regard ceding medical decision‐making to black box systems as contravening the profound moral responsibilities of clinicians. I argue, however, that opaque decisions are more common in medicine than critics realize. Moreover, as Aristotle noted over two millennia ago, when our knowledge of causal systems is incomplete and precarious—as it often is in medicine—the ability to explain how results are produced can be less important than the ability to produce such results and empirically verify their accuracy. (shrink)

This paper focuses on the use of ‘black box’ AI in medicine and asks whether the physician needs to disclose to patients that even the best AI comes with the risks of cyberattacks, systematic bias, and a particular type of mismatch between AI’s implicit assumptions and an individual patient’s background situation.Pacecurrent clinical practice, I argue that, under certain circumstances, these risks do need to be disclosed. Otherwise, the physician either vitiates a patient’s informed consent or violates a more general obligation (...) to warn him about potentially harmful consequences. To support this view, I argue, first, that the already widely accepted conditions in the evaluation of risks, i.e. the ‘nature’ and ‘likelihood’ of risks, speak in favour of disclosure and, second, that principled objections against the disclosure of these risks do not withstand scrutiny. Moreover, I also explain that these risks are exacerbated by pandemics like the COVID-19 crisis, which further emphasises their significance. (shrink)

In this paper, we first classify different types of second opinions and evaluate the ethical and epistemological implications of providing those in a clinical context. Second, we discuss the issue of how artificial intelligent could replace the human cognitive labour of providing such second opinion and find that several AI reach the levels of accuracy and efficiency needed to clarify their use an urgent ethical issue. Third, we outline the normative conditions of how AI may be used as second opinion (...) in clinical processes, weighing the benefits of its efficiency against concerns of responsibility attribution. Fourth, we provide a ‘rule of disagreement’ that fulfils these conditions while retaining some of the benefits of expanding the use of AI-based decision support systems in clinical contexts. This is because the rule of disagreement proposes to use AI as much as possible, but retain the ability to use human second opinions to resolve disagreements between AI and physician-in-charge. Fifth, we discuss some counterarguments. (shrink)

The use of machine learning systems for decision-support in healthcare may exacerbate health inequalities. However, recent work suggests that algorithms trained on sufficiently diverse datasets could in principle combat health inequalities. One concern about these algorithms is that their performance for patients in traditionally disadvantaged groups exceeds their performance for patients in traditionally advantaged groups. This renders the algorithmic decisions unfair relative to the standard fairness metrics in machine learning. In this paper, we defend the permissible use of affirmative algorithms; (...) that is, algorithms trained on diverse datasets that perform better for traditionally disadvantaged groups. Whilst such algorithmic decisions may be unfair, the fairness of algorithmic decisions is not the appropriate locus of moral evaluation. What matters is the fairness of final decisions, such as diagnoses, resulting from collaboration between clinicians and algorithms. We argue that affirmative algorithms can permissibly be deployed provided the resultant final decisions are fair. (shrink)

Several science studies scholars report instances of scientific “hype,” or sensationalized exaggeration, in journal articles, institutional press releases, and science journalism in a variety of fields (e.g., Caulfield and Condit 2012). Yet, how “hype” is being conceived varies. I will argue that hype is best understood as a particular kind of exaggeration, one that explicitly or implicitly exaggerates various positive aspects of science in ways that undermine the goals of science communication in a particular context. This account also makes clear (...) the ways that value judgments play a role in judgments of “hype,” which has implications for detecting and addressing this problem. (shrink)

Future weapons will make life-or-death decisions without a human in the loop. When such weapons inflict unwarranted harm, no one appears to be responsible. There seems to be a responsibility gap. I first reconstruct the argument for such responsibility gaps to then argue that this argument is not sound. The argument assumes that commanders have no control over whether autonomous weapons inflict harm. I argue against this assumption. Although this investigation concerns a specific case of autonomous weapons systems, I take (...) steps towards vindicating the more general idea that superiors can be morally responsible in virtue of being in command. (shrink)

Predictive algorithms are playing an increasingly prominent role in society, being used to predict recidivism, loan repayment, job performance, and so on. With this increasing influence has come an increasing concern with the ways in which they might be unfair or biased against individuals in virtue of their race, gender, or, more generally, their group membership. Many purported criteria of algorithmic fairness concern statistical relationships between the algorithm’s predictions and the actual outcomes, for instance requiring that the rate of false (...) positives be equal across the relevant groups. We might seek to ensure that algorithms satisfy all of these purported fairness criteria. But a series of impossibility results shows that this is impossible, unless base rates are equal across the relevant groups. What are we to make of these pessimistic results? I argue that none of the purported criteria, except for a calibration criterion, are necessary conditions for fairness, on the grounds that they can all be simultaneously violated by a manifestly fair and uniquely optimal predictive algorithm, even when base rates are equal. I conclude with some general reflections on algorithmic fairness. (shrink)

I outline a number of parallels between trust and distrust, emphasising the significance of situations in which both trust and distrust would be an imposition upon the (dis)trustee. I develop an account of both trust and distrust in terms of commitment, and argue that this enables us to understand the nature of trustworthiness. Note that this article is available open access on the journal website.

In recent years, a plethora of high-profile scientific publications has been reporting about machine learning algorithms outperforming clinicians in medical diagnosis or treatment recommendations. This has spiked interest in deploying relevant algorithms with the aim of enhancing decision-making in healthcare. In this paper, we argue that instead of straightforwardly enhancing the decision-making capabilities of clinicians and healthcare institutions, deploying machines learning algorithms entails trade-offs at the epistemic and the normative level. Whereas involving machine learning might improve the accuracy of medical (...) diagnosis, it comes at the expense of opacity when trying to assess the reliability of given diagnosis. Drawing on literature in social epistemology and moral responsibility, we argue that the uncertainty in question potentially undermines the epistemic authority of clinicians. Furthermore, we elucidate potential pitfalls of involving machine learning in healthcare with respect to paternalism, moral responsibility and fairness. At last, we discuss how the deployment of machine learning algorithms might shift the evidentiary norms of medical diagnosis. In this regard, we hope to lay the grounds for further ethical reflection of the opportunities and pitfalls of machine learning for enhancing decision-making in healthcare. (shrink)

Bioethics, Volume 36, Issue 2, Page 134-142, February 2022.

The term robustness is ubiquitous in modern Machine Learning (ML). However, its meaning varies depending on context and community. Researchers either focus on narrow technical definitions, such as adversarial robustness, natural distribution shifts, and performativity, or they simply leave open what exactly they mean by robustness. In this paper, we provide a conceptual analysis of the term robustness, with the aim to develop a common language, that allows us to weave together different strands of robustness research. We define robustness as (...) the relative stability of a robustness target with respect to specific interventions on a modifier. Our account captures the various sub-types of robustness that are discussed in the research literature, including robustness to distribution shifts, prediction robustness, or the robustness of algorithmic explanations. Finally, we delineate robustness from adjacent key concepts in ML, such as extrapolation, generalization, and uncertainty, and establish it as an independent epistemic concept. (shrink)

Mainstream epistemology is a highly theoretical and abstract enterprise. Traditional epistemologists rarely present their deliberations as critical to the practical problems of life, unless one supposes—as Hume, for example, did not—that skeptical worries should trouble us in our everyday affairs. But some issues in epistemology are both theoretically interesting and practically quite pressing. That holds of the problem to be discussed here: how laypersons should evaluate the testimony of experts and decide which of two or more rival experts is most (...) credible. It is of practical importance because in a complex, highly specialized world people are constantly confronted with situations in which, as comparative novices, they must turn to putative experts for intellectual guidance or assistance. It is of theoretical interest because the appropriate epistemic considerations are far from transparent; and it is not clear how far the problems lead to insurmountable skeptical quandaries. This paper does not argue for flat-out skepticism in this domain; nor, on the other hand, does it purport to resolve all pressures in the direction of skepticism. It is an exploratory paper, which tries to identify problems and examine some possible solutions, not to establish those solutions definitively. (shrink)

It is sometimes thought that randomized study group allocation is uniquely proficient at producing comparison groups that are evenly balanced for all confounding causes. Philosophers have argued that in real randomized controlled trials this balance assumption typically fails. But is the balance assumption an important ideal? I run a thought experiment, the CONFOUND study, to answer this question. I then suggest a new account of causal inference in ideal and real comparative group studies that helps clarify the roles of confounding (...) variables and randomization. 1Confounders and Causes2The Balance Assumption3The CONFOUND Study 3.1CONFOUND 13.2CONFOUND 24Disjunction C and the Ideal Study 4.1The ultimate ‘other cause’: C4.2The ideal comparative group study4.3Required conditions for causal inference5Confounders as Causes, Confounders as Correlates6Summary. (shrink)

The use of black box algorithms in medicine has raised scholarly concerns due to their opaqueness and lack of trustworthiness. Concerns about potential bias, accountability and responsibility, patient autonomy and compromised trust transpire with black box algorithms. These worries connect epistemic concerns with normative issues. In this paper, we outline that black box algorithms are less problematic for epistemic reasons than many scholars seem to believe. By outlining that more transparency in algorithms is not always necessary, and by explaining that (...) computational processes are indeed methodologically opaque to humans, we argue that the reliability of algorithms provides reasons for trusting the outcomes of medical artificial intelligence. To this end, we explain how computational reliabilism, which does not require transparency and supports the reliability of algorithms, justifies the belief that results of medical AI are to be trusted. We also argue that several ethical concerns remain with black box algorithms, even when the results are trustworthy. Having justified knowledge from reliable indicators is, therefore, necessary but not sufficient for normatively justifying physicians to act. This means that deliberation about the results of reliable algorithms is required to find out what is a desirable action. Thus understood, we argue that such challenges should not dismiss the use of black box algorithms altogether but should inform the way in which these algorithms are designed and implemented. When physicians are trained to acquire the necessary skills and expertise, and collaborate with medical informatics and data scientists, black box algorithms can contribute to improving medical care. (shrink)

There is considerable enthusiasm about the prospect that artificial intelligence (AI) will help to improve the safety and efficacy of health services and the efficiency of health systems. To realize this potential, however, AI systems will have to overcome structural problems in the culture and practice of medicine and the organization of health systems that impact the data from which AI models are built, the environments into which they will be deployed, and the practices and incentives that structure their development. (...) This perspective elaborates on some of these structural challenges and provides recommendations to address potential shortcomings. (shrink)

Various publications claim that medical AI systems perform as well, or better, than clinical experts. However, there have been very few controlled trials and the quality of existing studies has been called into question. There is growing concern that existing studies overestimate the clinical benefits of AI systems. This has led to calls for more, and higher-quality, randomized controlled trials of medical AI systems. While this a welcome development, AI RCTs raise novel methodological challenges that have seen little discussion. We (...) discuss some of the challenges arising in the context of AI RCTs and make some suggestions for how to meet them. (shrink)