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 paper shows that the law, in subtle ways, may set hitherto unrecognized incentives for the adoption of explainable machine learning applications. In doing so, we make two novel contributions. First, on the legal side, we show that to avoid liability, professional actors, such as doctors and managers, may soon be legally compelled to use explainable ML models. We argue that the importance of explainability reaches far beyond data protection law, and crucially influences questions of contractual and tort liability for (...) the use of ML models. To this effect, we conduct two legal case studies, in medical and corporate merger applications of ML. As a second contribution, we discuss the trade-off between accuracy and explainability and demonstrate the effect in a technical case study in the context of spam classification. (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)

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)

The debate about the ethical implications of Artificial Intelligence dates from the 1960s :741–742, 1960; Wiener in Cybernetics: or control and communication in the animal and the machine, MIT Press, New York, 1961). However, in recent years symbolic AI has been complemented and sometimes replaced by Neural Networks and Machine Learning techniques. This has vastly increased its potential utility and impact on society, with the consequence that the ethical debate has gone mainstream. Such a debate has primarily focused on principles—the (...) ‘what’ of AI ethics —rather than on practices, the ‘how.’ Awareness of the potential issues is increasing at a fast rate, but the AI community’s ability to take action to mitigate the associated risks is still at its infancy. Our intention in presenting this research is to contribute to closing the gap between principles and practices by constructing a typology that may help practically-minded developers apply ethics at each stage of the Machine Learning development pipeline, and to signal to researchers where further work is needed. The focus is exclusively on Machine Learning, but it is hoped that the results of this research may be easily applicable to other branches of AI. The article outlines the research method for creating this typology, the initial findings, and provides a summary of future research needs. (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)

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)

The concept of meaningful work has recently received increased attention in philosophy and other disciplines. However, the impact of the increasing robotization of the workplace on meaningful work has received very little attention so far. Doing work that is meaningful leads to higher job satisfaction and increased worker well-being, and some argue for a right to access to meaningful work. In this paper, we therefore address the impact of robotization on meaningful work. We do so by identifying five key aspects (...) of meaningful work: pursuing a purpose, social relationships, exercising skills and self-development, self-esteem and recognition, and autonomy. For each aspect, we analyze how the introduction of robots into the workplace may diminish or enhance the meaningfulness of work. We also identify a few ethical issues that emerge from our analysis. We conclude that robotization of the workplace can have both significant negative and positive effects on meaningful work. Our findings about ways in which robotization of the workplace can be a threat or opportunity for meaningful work can serve as the basis for ethical arguments for how to—and how not to—implement robots into workplaces. (shrink)

Many of the computing systems programmed using Machine Learning are opaque: it is difficult to know why they do what they do or how they work. Explainable Artificial Intelligence aims to develop analytic techniques that render opaque computing systems transparent, but lacks a normative framework with which to evaluate these techniques’ explanatory successes. The aim of the present discussion is to develop such a framework, paying particular attention to different stakeholders’ distinct explanatory requirements. Building on an analysis of “opacity” from (...) philosophy of science, this framework is modeled after accounts of explanation in cognitive science. The framework distinguishes between the explanation-seeking questions that are likely to be asked by different stakeholders, and specifies the general ways in which these questions should be answered so as to allow these stakeholders to perform their roles in the Machine Learning ecosystem. By applying the normative framework to recently developed techniques such as input heatmapping, feature-detector visualization, and diagnostic classification, it is possible to determine whether and to what extent techniques from Explainable Artificial Intelligence can be used to render opaque computing systems transparent and, thus, whether they can be used to solve the Black Box Problem. (shrink)

Artificial intelligence is set to transform healthcare. Key ethical issues to emerge with this transformation encompass the accountability and transparency of the decisions made by AI-based systems, the potential for group harms arising from algorithmic bias and the professional roles and integrity of clinicians. These concerns must be balanced against the imperatives of generating public benefit with more efficient healthcare systems from the vastly higher and accurate computational power of AI. In weighing up these issues, this paper applies the deliberative (...) balancing approach of the Ethics Framework for Big Data in Health and Research. The analysis applies relevant values identified from the framework to demonstrate how decision-makers can draw on them to develop and implement AI-assisted support systems into healthcare and clinical practice ethically and responsibly. Please refer to Xafis et al. in this special issue of the Asian Bioethics Review for more information on how this framework is to be used, including a full explanation of the key values involved and the balancing approach used in the case study at the end of this paper. (shrink)

The usefulness of machine learning algorithms has led to their widespread adoption prior to the development of a conceptual framework for making sense of them. One common response to this situation is to say that machine learning suffers from a “black box problem.” That is, machine learning algorithms are “opaque” to human users, failing to be “interpretable” or “explicable” in terms that would render categorization procedures “understandable.” The purpose of this paper is to challenge the widespread agreement about the existence (...) and importance of a black box problem. The first section argues that “interpretability” and cognates lack precise meanings when applied to algorithms. This makes the concepts difficult to use when trying to solve the problems that have motivated the call for interpretability. Furthermore, since there is no adequate account of the concepts themselves, it is not possible to assess whether particular technical features supply formal definitions of those concepts. The second section argues that there are ways of being a responsible user of these algorithms that do not require interpretability. In many cases in which a black box problem is cited, interpretability is a means to a further end such as justification or non-discrimination. Since addressing these problems need not involve something that looks like an “interpretation” of an algorithm, the focus on interpretability artificially constrains the solution space by characterizing one possible solution as the problem itself. Where possible, discussion should be reformulated in terms of the ends of interpretability. (shrink)

In an analysis of artificially intelligent systems for medical diagnostics and treatment planning we argue that patients should be able to exercise a right to withdraw from AI diagnostics and treatment planning for reasons related to (1) the physician’s role in the patients’ formation of and acting on personal preferences and values, (2) the bias and opacity problem of AI systems, and (3) rational concerns about the future societal effects of introducing AI systems in the health care sector.

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)

As we increasingly delegate decision-making to algorithms, whether directly or indirectly, important questions emerge in circumstances where those decisions have direct consequences for individual rights and personal opportunities, as well as for the collective good. A key problem for policymakers is that the social implications of these new methods can only be grasped if there is an adequate comprehension of their general technical underpinnings. The discussion here focuses primarily on the case of enforcement decisions in the criminal justice system, but (...) draws on similar situations emerging from other algorithms utilised in controlling access to opportunities, to explain how machine learning works and, as a result, how decisions are made by modern intelligent algorithms or 'classifiers'. It examines the key aspects of the performance of classifiers, including how classifiers learn, the fact that they operate on the basis of correlation rather than causation, and that the term 'bias' in machine learning has a different meaning to common usage. An example of a real world 'classifier', the Harm Assessment Risk Tool (HART), is examined, through identification of its technical features: the classification method, the training data and the test data, the features and the labels, validation and performance measures. Four normative benchmarks are then considered by reference to HART: (a) prediction accuracy (b) fairness and equality before the law (c) transparency and accountability (d) informational privacy and freedom of expression, in order to demonstrate how its technical features have important normative dimensions that bear directly on the extent to which the system can be regarded as a viable and legitimate support for, or even alternative to, existing human decision-makers. (shrink)

I analyse an argument according to which medical artificial intelligence represents a threat to patient autonomy—recently put forward by Rosalind McDougall in the Journal of Medical Ethics. The argument takes the case of IBM Watson for Oncology to argue that such technologies risk disregarding the individual values and wishes of patients. I find three problems with this argument: it confuses AI with machine learning; it misses machine learning’s potential for personalised medicine through big data; it fails to distinguish between evidence-based (...) advice and decision-making within healthcare. I conclude that how much and which tasks we should delegate to machine learning and other technologies within healthcare and beyond is indeed a crucial question of our time, but in order to answer it, we must be careful in analysing and properly distinguish between the different systems and different delegated tasks. (shrink)

While organizations today make extensive use of complex algorithms, the notion of algorithmic accountability remains an elusive ideal due to the opacity and fluidity of algorithms. In this article, we develop a framework for managing algorithmic accountability that highlights three interrelated dimensions: reputational concerns, engagement strategies, and discourse principles. The framework clarifies that accountability processes for algorithms are driven by reputational concerns about the epistemic setup, opacity, and outcomes of algorithms; that the way in which organizations practically engage with emergent (...) expectations about algorithms may be manipulative, adaptive, or moral; and that when accountability relationships are heavily burdened by the opacity and fluidity of complex algorithmic systems, the emphasis of engagement should shift to a rational communication process through which a continuous and tentative assessment of the development, workings, and consequences of algorithms can be achieved over time. The degree to which such engagement is, in fact, rational can be assessed based on four discourse-ethical principles of participation, comprehension, multivocality, and responsiveness. We conclude that the framework may help organizations and their environments to jointly work toward greater accountability for complex algorithms. It may further help organizations in reputational positioning surrounding accountability issues. The discourse-ethical principles introduced in this article are meant to elevate these positioning contests to extend beyond mere adaption or compliance and help guide organizations to find moral and forward-looking solutions to accountability issues. (shrink)

Algorithms can now identify patterns and correlations in the (big) datasets, and predict outcomes based on those identified patterns and correlations with the use of machine learning techniques and big data, decisions can then be made by algorithms themselves in accordance with the predicted outcomes. Yet, algorithms can inherit questionable values from the datasets and acquire biases in the course of (machine) learning, and automated algorithmic decision-making makes it more difficult for people to see algorithms as biased. While researchers have (...) taken the problem of algorithmic bias seriously, but the current discussion on algorithmic fairness tends to conceptualize ‘fairness’ in algorithmic fairness primarily as a technical issue and attempts to implement pre-existing ideas of ‘fairness’ into algorithms. In this paper, I show that such a view of algorithmic fairness as technical issue is unsatisfactory for the type of problem algorithmic fairness presents. Since decisions on fairness measure and the related techniques for algorithms essentially involve choices between competing values, ‘fairness’ in algorithmic fairness should be conceptualized first and foremost as a political issue, and it should be (re)solved by democratic communication. The aim of this paper, therefore, is to explicitly reconceptualize algorithmic fairness as a political question and suggest the current discussion of algorithmic fairness can be strengthened by adopting the accountability for reasonableness framework. (shrink)

Modern health data practices come with many practical uncertainties. In this paper, I argue that data subjects’ trust in the institutions and organizations that control their data, and their ability to know their own moral obligations in relation to their data, are undermined by significant uncertainties regarding the what, how, and who of mass data collection and analysis. I conclude by considering how proposals for managing situations of high uncertainty might be applied to this problem. These emphasize increasing organizational flexibility, (...) knowledge, and capacity, and reducing hazard. (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)

Interactions between an intelligent software agent and a human user are ubiquitous in everyday situations such as access to information, entertainment, and purchases. In such interactions, the ISA mediates the user’s access to the content, or controls some other aspect of the user experience, and is not designed to be neutral about outcomes of user choices. Like human users, ISAs are driven by goals, make autonomous decisions, and can learn from experience. Using ideas from bounded rationality, we frame these interactions (...) as instances of an ISA whose reward depends on actions performed by the user. Such agents benefit by steering the user’s behaviour towards outcomes that maximise the ISA’s utility, which may or may not be aligned with that of the user. Video games, news recommendation aggregation engines, and fitness trackers can all be instances of this general case. Our analysis facilitates distinguishing various subcases of interaction, as well as second-order effects that might include the possibility for adaptive interfaces to induce behavioural addiction, and/or change in user belief. We present these types of interaction within a conceptual framework, and review current examples of persuasive technologies and the issues that arise from their use. We argue that the nature of the feedback commonly used by learning agents to update their models and subsequent decisions could steer the behaviour of human users away from what benefits them, and in a direction that can undermine autonomy and cause further disparity between actions and goals as exemplified by addictive and compulsive behaviour. We discuss some of the ethical, social and legal implications of this technology and argue that it can sometimes exploit and reinforce weaknesses in human beings. (shrink)

Personal AI assistants are now nearly ubiquitous. Every leading smartphone operating system comes with a personal AI assistant that promises to help you with basic cognitive tasks: searching, planning, messaging, scheduling and so on. Usage of such devices is effectively a form of algorithmic outsourcing: getting a smart algorithm to do something on your behalf. Many have expressed concerns about this algorithmic outsourcing. They claim that it is dehumanising, leads to cognitive degeneration, and robs us of our freedom and autonomy. (...) Some people have a more subtle view, arguing that it is problematic in those cases where its use may degrade important interpersonal virtues. In this article, I assess these objections to the use of AI assistants. I will argue that the ethics of their use is complex. There are no quick fixes or knockdown objections to the practice, but there are some legitimate concerns. By carefully analysing and evaluating the objections that have been lodged to date, we can begin to articulate an ethics of personal AI use that navigates those concerns. In the process, we can locate some paradoxes in our thinking about outsourcing and technological dependence, and we can think more clearly about what it means to live a good life in the age of smart machines. (shrink)

Algorithms silently structure our lives. Algorithms can determine whether someone is hired, promoted, offered a loan, or provided housing as well as determine which political ads and news articles consumers see. Yet, the responsibility for algorithms in these important decisions is not clear. This article identifies whether developers have a responsibility for their algorithms later in use, what those firms are responsible for, and the normative grounding for that responsibility. I conceptualize algorithms as value-laden, rather than neutral, in that algorithms (...) create moral consequences, reinforce or undercut ethical principles, and enable or diminish stakeholder rights and dignity. In addition, algorithms are an important actor in ethical decisions and influence the delegation of roles and responsibilities within these decisions. As such, firms should be responsible not only for the value-laden-ness of an algorithm but also for designing who-does-what within the algorithmic decision. As such, firms developing algorithms are accountable for designing how large a role individual will be permitted to take in the subsequent algorithmic decision. Counter to current arguments, I find that if an algorithm is designed to preclude individuals from taking responsibility within a decision, then the designer of the algorithm should be held accountable for the ethical implications of the algorithm in use. (shrink)

This paper explores and rehabilitates the value of decisional privacy as a conceptual tool, complementary to informational privacy, for critiquing personalized choice architectures employed by self-tracking technologies. Self-tracking technologies are promoted and used as a means to self-improvement. Based on large aggregates of personal data and the data of other users, self-tracking technologies offer personalized feedback that nudges the user into behavioral change. The real-time personalization of choice architectures requires continuous surveillance and is a very powerful technology, recently coined as (...) “hypernudging.” While users celebrate the increased personalization of their coaching devices, “hypernudging” technologies raise concerns about manipulation. This paper addresses that intuition by claiming that decisional privacy is at stake. It thus counters the trend to solely focus on informational privacy when evaluating information and communication technologies. It proposes that decisional privacy and informational privacy are often part of a mutually reinforcing dynamic. Hypernudging is used as a key example to illustrate that the two dimensions should not be treated separately. Hypernudging self-tracking technologies compromise autonomy because they violate informational and decisional privacy. In order to effectively judge whether technologies that use hypernudges empower users, we need both privacy dimensions as conceptual tools. (shrink)

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

This paper explores the concept of digital modernity, the extension of narratives of modernity with the special affordances of digital networked technology. Digital modernity produces a new narrative which can be taken in many ways: to be descriptive of reality; a teleological account of an inexorable process; or a normative account of an ideal sociotechnical state. However, it is understood that narratives of digital modernity help shape reality via commercial and political decision-makers, and examples are given from the politics and (...) society of the United Kingdom. The paper argues that digital modernity has two dimensions, of progression through time and progression through space, and these two dimensions can be in contradiction. Contradictions can also be found between ideas of digital modernity and modernity itself, and also between digital modernity and some of the basic pre-modern concepts that underlie the whole technology industry. Therefore, digital modernity may not be a sustainable goal for technology development. (shrink)

This discussion note does three things: it explains the notion of ‘legal protection by design’ in relation to data-driven infrastructures that form the backbone of our new ‘onlife world’, it explains how the notion of ‘by design’ relates to the relational nature of what an environment affords its inhabitants, referring to the work of James Gibson, and it explains how this affects our understanding of human capabilities in relation to the affordances of changing environments. Finally, this brief note argues that (...) ‘safer by design’ in the case of nanotechnology will require legal protection by design to make sure that human capabilities are reinvented and sustained in nano-technical environments. (shrink)

Data science is not simply a method but an organising idea. Commitment to the new paradigm overrides concerns caused by collateral damage, and only a counterculture can constitute an effective critique. Understanding data science requires an appreciation of what algorithms actually do; in particular, how machine learning learns. The resulting ‘insight through opacity’ drives the observable problems of algorithmic discrimination and the evasion of due process. But attempts to stem the tide have not grasped the nature of data science as (...) both metaphysical and machinic. Data science strongly echoes the neoplatonism that informed the early science of Copernicus and Galileo. It appears to reveal a hidden mathematical order in the world that is superior to our direct experience. The new symmetry of these orderings is more compelling than the actual results. Data science does not only make possible a new way of knowing but acts directly on it; by converting predictions to pre-emptions, it becomes a machinic metaphysics. The people enrolled in this apparatus risk an abstraction of accountability and the production of ‘thoughtlessness’. Susceptibility to data science can be contested through critiques of science, especially standpoint theory, which opposes the ‘view from nowhere’ without abandoning the empirical methods. But a counterculture of data science must be material as well as discursive. Karen Barad’s idea of agential realism can reconfigure data science to produce both non-dualistic philosophy and participatory agency. An example of relevant praxis points to the real possibility of ‘machine learning for the people’. (shrink)

The internet of things is increasingly spreading into the domain of medical and social care. Internet-enabled devices for monitoring and managing the health and well-being of users outside of traditional medical institutions have rapidly become common tools to support healthcare. Health-related internet of things (H-IoT) technologies increasingly play a key role in health management, for purposes including disease prevention, real-time tele-monitoring of patient’s functions, testing of treatments, fitness and well-being monitoring, medication dispensation, and health research data collection. H-IoT promises many (...) benefits for health and healthcare. However, it also raises a host of ethical problems stemming from the inherent risks of Internet enabled devices, the sensitivity of health-related data, and their impact on the delivery of healthcare. This paper maps the main ethical problems that have been identified by the relevant literature and identifies key themes in the on-going debate on ethical problems concerning H-IoT. (shrink)

The ever-increasing application of algorithms to decision-making in a range of social contexts has prompted demands for algorithmic accountability. Accountable decision-makers must provide their decision-subjects with justifications for their automated system’s outputs, but what kinds of broader principles should we expect such justifications to appeal to? Drawing from political philosophy, I present an account of algorithmic accountability in terms of the democratic ideal of ‘public reason’. I argue that situating demands for algorithmic accountability within this justificatory framework enables us to (...) better articulate their purpose and assess the adequacy of efforts toward them. (shrink)

Mature information societies are characterised by mass production of data that provide insight into human behaviour. Analytics has arisen as a practice to make sense of the data trails generated through interactions with networked devices, platforms and organisations. Persistent knowledge describing the behaviours and characteristics of people can be constructed over time, linking individuals into groups or classes of interest to the platform. Analytics allows for a new type of algorithmically assembled group to be formed that does not necessarily align (...) with classes or attributes already protected by privacy and anti-discrimination law or addressed in fairness- and discrimination-aware analytics. Individuals are linked according to offline identifiers and shared behavioural identity tokens, allowing for predictions and decisions to be taken at a group rather than individual level. This article examines the ethical significance of such ad hoc groups in analytics and argues that the privacy interests of algorithmically assembled groups in inviolate personality must be recognised alongside individual privacy rights. Algorithmically grouped individuals have a collective interest in the creation of information about the group, and actions taken on its behalf. Group privacy is proposed as a third interest to balance alongside individual privacy and social, commercial and epistemic benefits when assessing the ethical acceptability of analytics platforms. (shrink)

Recent work in explainable artificial intelligence (XAI) attempts to render opaque AI systems understandable through a divide-and-conquer strategy. However, this fails to illuminate how trained AI systems work as a whole. Precisely this kind of functional understanding is needed, though, to satisfy important societal desiderata such as safety. To remedy this situation, we argue, AI researchers should seek mechanistic interpretability, viz. apply coordinated discovery strategies familiar from the life sciences to uncover the functional organisation of complex AI systems. Additionally, theorists (...) should accommodate for the unique costs and benefits of such strategies in their portrayals of XAI research. (shrink)

In the ethics of algorithms, a specifically epistemological analysis is rarely undertaken in order to gain a critique (or a defense) of the handling of or trust in medical black box algorithms (BBAs). This article aims to begin to fill this research gap. Specifically, the thesis is examined according to which such algorithms are regarded as epistemic authorities (EAs) and that the results of a medical algorithm must completely replace other convictions that patients have (preemptionism). If this were true, it (...) would be a reason to distrust medical BBAs. First, the author describes what EAs are and why BBAs can be considered EAs. Then, preemptionism will be outlined and criticized as an answer to the question of how to deal with an EA. The discussion leads to some requirements for dealing with a BBA as an EA. (shrink)

Artificial intelligence (AI) systems have demonstrated impressive performance across a variety of clinical tasks. However, notoriously, sometimes these systems are “black boxes.” The initial response in the literature was a demand for “explainable AI.” However, recently, several authors have suggested that making AI more explainable or “interpretable” is likely to be at the cost of the accuracy of these systems and that prioritizing interpretability in medical AI may constitute a “lethal prejudice.” In this paper, we defend the value of interpretability (...) in the context of the use of AI in medicine. Clinicians may prefer interpretable systems over more accurate black boxes, which in turn is sufficient to give designers of AI reason to prefer more interpretable systems in order to ensure that AI is adopted and its benefits realized. Moreover, clinicians may be justified in this preference. Achieving the downstream benefits from AI is critically dependent on how the outputs of these systems are interpreted by physicians and patients. A preference for the use of highly accurate black box AI systems, over less accurate but more interpretable systems, may itself constitute a form of lethal prejudice that may diminish the benefits of AI to—and perhaps even harm—patients. (shrink)

In information societies, operations, decisions and choices previously left to humans are increasingly delegated to algorithms, which may advise, if not decide, about how data should be interpreted and what actions should be taken as a result. More and more often, algorithms mediate social processes, business transactions, governmental decisions, and how we perceive, understand, and interact among ourselves and with the environment. Gaps between the design and operation of algorithms and our understanding of their ethical implications can have severe consequences (...) affecting individuals as well as groups and whole societies. This paper makes three contributions to clarify the ethical importance of algorithmic mediation. It provides a prescriptive map to organise the debate. It reviews the current discussion of ethical aspects of algorithms. And it assesses the available literature in order to identify areas requiring further work to develop the ethics of algorithms. (shrink)

Technology could be used to improve morality but it could do so in different ways. Some technologies could augment and enhance moral behaviour externally by using external cues and signals to push and pull us towards morally appropriate behaviours. Other technologies could enhance moral behaviour internally by directly altering the way in which the brain captures and processes morally salient information or initiates moral action. The question is whether there is any reason to prefer one method over the other? In (...) this article, I argue that there is. Specifically, I argue that internal moral enhancement is likely to be preferable to external moral enhancement, when it comes to the legitimacy of political decision-making processes. In fact, I go further than this and argue that the increasingly dominant forms of external moral enhancement may already be posing a significant threat to political legitimacy, one that we should try to address. Consequently, research and development of internal moral enhancements should be prioritised as a political project. (shrink)

It is widely believed that a conservative moral outlook is opposed to biomedical forms of human enhancement. In this paper, I argue that this widespread belief is incorrect. Using Cohen’s evaluative conservatism as my starting point, I argue that there are strong conservative reasons to prioritise the development of biomedical enhancements. In particular, I suggest that biomedical enhancement may be essential if we are to maintain our current evaluative equilibrium (i.e. the set of values that undergird and permeate our current (...) political, economic, and personal lives) against the threats to that equilibrium posed by external, non-biomedical forms of enhancement. I defend this view against modest conservatives who insist that biomedical enhancements pose a greater risk to our current evaluative equilibrium, and against those who see no principled distinction between the forms of human enhancement. (shrink)

Released mid-November 2022, Galactica is a set of six large language models (LLMs) of different sizes (from 125 M to 120B parameters) designed by Meta AI to achieve the ultimate ambition of “a single neural network for powering scientific tasks”, according to its accompanying whitepaper. It aims to carry out knowledge-intensive tasks, such as publication summarization, information ordering and protein annotation. However, just a few days after the release, Meta had to pull back the demo due to the strong hallucinatory (...) tendencies or underwhelming performances of the model. This article aims to study, through a critical threefold argument, the potential impacts of LLMs once deployed in the scientific value chain. Our first argument is a technical one. By examining the technicity of Galactica, it is possible to explain the descripancies between its promotional corporate discourse and abysmal outputs. Second, by going back to debates in both computer science and computational philosophy on the automation of abduction, we argue from the epistemological front that LLMs indeed cannot produce strong abductions and, therefore, claims about the automation of hypothesis generation remains chambering. Finally, our third argument is a sociological one. By conceptualizing the scientific field through Nancy Katherine Hayles’ cognitive assemblage theory, we aim to outline the potential steering of science by LLMs, mainly through information ordering. The core of our argument rests on the assertion that excessive control on information risks contravening a certain serendipitous aspect inherent to scientific discoveries. (shrink)

As opaque algorithmic systems take up a larger and larger role in shaping our lives, calls for explainability in various algorithmic systems have increased. Many moral and political philosophers have sought to vindicate these calls for explainability by developing theories on which decision-subjects—that is, individuals affected by decisions—have a moral right to the explanation of the systems that affect them. Existing theories tend to suggest that the right to explanation arises solely in virtue of facts about how decision-subjects are affected (...) by opaque systems. But this ignores an important part of the normative landscape: sometimes, facts about a decision-maker’s rights and freedoms can make a difference to whether she owes anyone explanations of her decisions. I illustrate the point by calling attention to examples of artistic, romantic, and supererogatory altruistic decisions; in at least some cases of this sort, even highly influential decisions can legitimately be kept entirely private and opaque. (shrink)

The ongoing explosion of interest in artificial intelligence is fueled in part by recently developed techniques in machine learning. Those techniques allow automated systems to process huge amounts of data, utilizing mathematical methods that depart from traditional statistical approaches, and resulting in impressive advancements in our ability to make predictions and uncover correlations across a host of interesting domains. But as is now widely discussed, the way that those systems arrive at their outputs is often opaque, even to the experts (...) who design and deploy them. Is it morally problematic to make use of opaque automated methods when making high-stakes decisions, like whether to issue a loan to an applicant, or whether to approve a parole request? Many scholars answer in the affirmative. However, there is no widely accepted explanation for why transparent systems are morally preferable to opaque systems. We argue that the use of automated decision-making systems sometimes violates duties of consideration that are owed by decision-makers to decision-subjects, duties that are both epistemic and practical in character. Violations of that kind generate a weighty consideration against the use of opaque decision systems. In the course of defending our approach, we show that it is able to address three major challenges sometimes leveled against attempts to defend the moral import of transparency in automated decision-making. (shrink)

Machine learning-based AI algorithms lack transparency. In this article, I offer an interpretation of AI’s explainability problem and highlight its ethical saliency. I try to make the case for the legal enforcement of a strong explainability requirement: human organizations which decide to automate decision-making should be legally obliged to demonstrate the capacity to explain and justify the algorithmic decisions that have an impact on the wellbeing, rights, and opportunities of those affected by the decisions. This legal duty can be derived (...) from the demands of Rawlsian public reason. In the second part of the paper, I try to show that the argument from the limitations of human cognition fails to get AI off the hook of public reason. Against a growing trend in AI ethics, my main argument is that the analogy between human minds and artificial neural networks fails because it suffers from an atomistic bias which makes it blind to the social and institutional dimension of human reasoning processes. I suggest that developing interpretive AI algorithms is not the only possible answer to the explainability problem; social and institutional answers are also available and in many cases more trustworthy than techno-scientific ones. (shrink)

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

Deep neural networks are said to be opaque, impeding the development of safe and trustworthy artificial intelligence, but where this opacity stems from is less clear. What are the sufficient properties for neural network opacity? Here, I discuss five common properties of deep neural networks and two different kinds of opacity. Which of these properties are sufficient for what type of opacity? I show how each kind of opacity stems from only one of these five properties, and then discuss to (...) what extent the two kinds of opacity can be mitigated by explainability methods. (shrink)

Many AI systems that make important decisions are black boxes: how they function is opaque even to their developers. This is due to their high complexity and to the fact that they are trained rather than programmed. Efforts to alleviate the opacity of black box systems are typically discussed in terms of transparency, interpretability, and explainability. However, there is little agreement about what these key concepts mean, which makes it difficult to adjudicate the success or promise of opacity alleviation methods. (...) I argue for a unified account of these key concepts that treats the concept of understanding as fundamental. This allows resources from the philosophy of science and the epistemology of understanding to help guide opacity alleviation efforts. A first significant benefit of this understanding account is that it defuses one of the primary, in-principle objections to post hoc explainable AI (XAI) methods. This “rationalization objection” argues that XAI methods provide mere rationalizations rather than genuine explanations. This is because XAI methods involve using a separate “explanation” system to approximate the original black box system. These explanation systems function in a completely different way than the original system, yet XAI methods make inferences about the original system based on the behavior of the explanation system. I argue that, if we conceive of XAI methods as idealized scientific models, this rationalization worry is dissolved. Idealized scientific models misrepresent their target phenomena, yet are capable of providing significant and genuine understanding of their targets. (shrink)

In this paper, I reflect on the puzzle that machine learning presents to social theory to develop an account of its distinct impact on social reality. I start by presenting how machine learning has presented a challenge to social theory as a research subject comprising both familiar and alien characteristics (1.). Taking this as an occasion for theoretical inquiry, I then propose a conceptual framework to investigate how algorithmic models of social phenomena relate to social reality and what their stochastic (...) mode of operation entails in terms of their sociality (2.). Analyzed through a theoretical lens that relies on central tenets of sociological systems theory, I find that machine learning implies a distinct epistemic transformation, based on how algorithmic modeling techniques process meaning as represented in data embedded in vector space. Building on this characterization, I introduce my conceptualization of stochastic technology as distinct from mechanistic technologies that rely on causal fixation (3.). Based on this understanding, I suggest that real-world applications of machine learning are often characterized by a constitutive tension between the stochastic properties of their outputs and the ways in which they are put to use in practice. Focussing on the large language models LaMDA and ChatGPT, I examine the epistemological implications of LLMs to account for the confusion of correlation and causality as the root of this tension. Next, I illustrate my theoretical conception by way of discussing an essay on image models by German media artist Hito Steyerl (4.). Following a critical reflection on Steyerl's characterization of Stable Diffusion as a “white box ”, I finally propose to conceive ofmachine learning-based technologies as stochastic contingency machines that transform social indeterminacy into contingent observations of social phenomena (5.) In this perspective, machine learning constitutes an epistemic technology that operates on meaning as extractable from data by means of algorithmic data modeling techniques to produce stochastic accounts of social reality. (shrink)

Large language models (LLMs) such as ChatGPT present immense opportunities, but without proper training for users (and potentially oversight), they carry risks of misuse as well. We argue that current approaches focusing predominantly on transparency and explainability fall short in addressing the diverse needs and concerns of various user groups. We highlight the limitations of existing methodologies and propose a framework anchored on user-centric guidelines. In particular, we argue that LLM users should be given guidelines on what tasks LLMs can (...) do well and which they cannot, which tasks require further guidance or refinement by the user, and context-specific heuristics. We further argue that (some) users should be taught to refine and elaborate adequate prompts, be provided with good procedures for prompt iteration, and be taught efficient ways to verify outputs. We suggest that for users, shifting away from looking at the technology itself, but rather looking at the usage of it within contextualized sociotechnical systems, can help solve many issues related to LLMs. We further emphasize the role of real-world case studies in shaping these guidelines, ensuring they are grounded in practical, applicable strategies. Like any technology, risks of misuse can be managed through education, regulation, and responsible development. (shrink)

Artificial intelligence (AI) has long been recognised as a challenge to responsibility. Much of this discourse has been framed around robots, such as autonomous weapons or self-driving cars, where we arguably lack control over a machine’s behaviour and therefore struggle to identify an agent that can be held accountable. However, most of today’s AI is based on machine-learning technology that does not act on its own, but rather serves as a decision-support tool, automatically analysing data to help human agents make (...) better decisions. I argue that decision-support tools pose a challenge to responsibility that goes beyond the familiar problem of finding someone to blame or punish for the behaviour of agent-like systems. Namely, they pose a problem for what we might call “decision ownership”: they make it difficult to identify human agents to whom we can attribute value-judgements that are reflected in decisions. Drawing on recent philosophical literature on responsibility and its various facets, I argue that this is primarily a problem of attributability rather than of accountability. This particular responsibility problem comes in different forms and degrees, most obviously when an AI provides direct recommendations for actions, but also, less obviously, when it provides mere descriptive information on the basis of which a decision is made. (shrink)

A central goal of research in explainable artificial intelligence (XAI) is to facilitate human understanding. However, understanding is an elusive concept that is difficult to target. In this paper, we argue that a useful way to conceptualize understanding within the realm of XAI is via certain human abilities. We present four criteria for a useful conceptualization of understanding in XAI and show that these are fulfilled by an abilities-based approach: First, thinking about understanding in terms of specific abilities is motivated (...) by research from numerous disciplines involved in XAI. Second, an abilities-based approach is highly versatile and can capture different forms of understanding important in XAI application contexts. Third, abilities can be operationalized for empirical studies. Fourth, abilities can be used to clarify the link between explainability, understanding, and societal desiderata concerning AI, like fairness and trustworthiness. Conceptualizing understanding as abilities can therefore support interdisciplinary collaboration among XAI researchers, provide practical benefit across diverse XAI application contexts, facilitate the development and evaluation of explainability approaches, and contribute to satisfying the societal desiderata of different stakeholders concerning AI systems. (shrink)

This paper delves into the use of machine learning-based systems in decision-making processes and its implications for moral responsibility as traditionally defined. It focuses on the emergence of responsibility gaps and examines proposed strategies to address them. The paper aims to provide an introductory and comprehensive overview of the ongoing debate surrounding moral responsibility in automated decision-making. By thoroughly examining these issues, we seek to contribute to a deeper understanding of the implications of AI integration in society.