The paper presents a framework to realise “meaningful human control” over Automated Driving Systems. The framework is based on an original synthesis of the results of the multidisciplinary research project “Meaningful Human Control over Automated Driving Systems” lead by a team of engineers, philosophers, and psychologists at Delft University of the Technology from 2017 to 2021. Meaningful human control aims at protecting safety and reducing responsibility gaps. The framework is based on the core assumption that human persons and institutions, not (...) hardware and software and their algorithms, should remain ultimately—though not necessarily directly—in control of, and thus morally responsible for, the potentially dangerous operation of driving in mixed traffic. We propose an Automated Driving System to be under meaningful human control if it behaves according to the relevant reasons of the relevant human actors (tracking), and that any potentially dangerous event can be related to a human actor (tracing). We operationalise the requirements for meaningful human control through multidisciplinary work in philosophy, behavioural psychology and traffic engineering. The tracking condition is operationalised via a proximal scale of reasons and the tracing condition via an evaluation cascade table. We review the implications and requirements for the behaviour and skills of human actors, in particular related to supervisory control and driver education. We show how the evaluation cascade table can be applied in concrete engineering use cases in combination with the definition of core components to expose deficiencies in traceability, thereby avoiding so-called responsibility gaps. Future research directions are proposed to expand the philosophical framework and use cases, supervisory control and driver education, real-world pilots and institutional embedding. (shrink)

The application of artificial intelligence (AI) to judicial decision-making has already begun in many jurisdictions around the world. While AI seems to promise greater fairness, access to justice, and legal certainty, issues of discrimination and transparency have emerged and put liberal democratic principles under pressure, most notably in the context of bail decisions. Despite this, there has been no systematic analysis of the risks to liberal democratic values from implementing AI into judicial decision-making. This article sets out to fill this (...) void by identifying and engaging with challenges arising from artificial judicial decision-making, focusing on three pillars of liberal democracy, namely equal treatment of citizens, transparency, and judicial independence. Methodologically, the work takes a comparative perspective between human and artificial decision-making, using the former as a normative benchmark to evaluate the latter. -/- The chapter first argues that AI that would improve on equal treatment of citizens has already been developed, but not yet adopted. Second, while the lack of transparency in AI decision-making poses severe risks which ought to be addressed, AI can also increase the transparency of options and trade-offs that policy makers face when considering the consequences of artificial judicial decision-making. Such transparency of options offers tremendous benefits from a democratic perspective. Third, the overall shift of power from human intuition to advanced AI may threaten judicial independence, and with it the separation of powers. While improvements regarding discrimination and transparency are available or on the horizon, it remains unclear how judicial independence can be protected, especially with the potential development of advanced artificial judicial intelligence (AAJI). Working out the political and legal infrastructure to reap the fruits of artificial judicial intelligence in a safe and stable manner should become a priority of future research in this area. (shrink)

Our understanding of what exactly needs protected against in order to safeguard a plausible construal of our ‘freedom of thought’ is changing. And this is because the recent influx of cognitive offloading and outsourcing—and the fast-evolving technologies that enable this—generate radical new possibilities for freedom-of-thought violating thought manipulation. This paper does three main things. First, I briefly overview how recent thinking in the philosophy of mind and cognitive science recognises—contrary to traditional Cartesian ‘internalist’ assumptions—ways in which our cognitive faculties, and (...) even our beliefs, can be materially realised by as well as stored non-biologically and extracranially. Second, and taking brain-computer interface technologies (BCIs) and the associated possibility of ‘extended’ beliefs as a reference point, I propose and defend a sufficient condition on freedom-of-thought violating (extended) thought manipulation. On the view proposed, the right not to have one’s thoughts or opinions manipulated is violated if one is (i) caused to acquire non-autonomous propositional attitudes (acquisition manipulation) or (ii) caused to have otherwise autonomous propositional attitudes non-autonomously eradicated (eradication manipulation). The implications of this view are then illustrated through four thought experiments, which map on to four distinct ways—what I call Type 1-Type 4 manipulation—in which, and with reference to the view defended, one’s freedom of thought is plausibly violated. (shrink)

The US is promoting a new vision of a “Good AI Society” through its recent AI Bill of Rights. This offers a promising vision of community-oriented equity unique amongst peer countries. However, it leaves the door open for potential rights violations. Furthermore, it may have some federal impact, but it is non-binding, and without concrete legislation, the private sector is likely to ignore it.

Artificial intelligence (AI) and robotics are digital technologies that will have significant impact on the development of humanity in the near future. They have raised fundamental questions about what we should do with these systems, what the systems themselves should do, what risks they involve, and how we can control these. - After the Introduction to the field (§1), the main themes (§2) of this article are: Ethical issues that arise with AI systems as objects, i.e., tools made and used (...) by humans. This includes issues of privacy (§2.1) and manipulation (§2.2), opacity (§2.3) and bias (§2.4), human-robot interaction (§2.5), employment (§2.6), and the effects of autonomy (§2.7). Then AI systems as subjects, i.e., ethics for the AI systems themselves in machine ethics (§2.8) and artificial moral agency (§2.9). Finally, the problem of a possible future AI superintelligence leading to a “singularity” (§2.10). We close with a remark on the vision of AI (§3). - For each section within these themes, we provide a general explanation of the ethical issues, outline existing positions and arguments, then analyse how these play out with current technologies and finally, what policy consequences may be drawn. (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)

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)

The General Data Protection Regulation (GDPR) of the EU confirms the protection of personal data as a fundamental human right and affords data subjects more control over the way their personal information is processed, shared, and analyzed. However, where data are processed by artificial intelligence (AI) algorithms, asserting control and providing adequate explanations is a challenge. Due to massive increases in computing power and big data processing, modern AI algorithms are too complex and opaque to be understood by most data (...) subjects. Articles 15 and 22 of the GDPR provide a modest regulatory framework for automated data processing by, among other things, mandating that data controllers inform data subjects about when it is being used, and its logic and ramifications. Nevertheless, due to the phrasing of the articles and the numerous exceptions they allow, doubts have arisen about their effectiveness. In this paper, we empirically evaluate the quality and effectiveness of AI disclosures as mandated by the GDPR. By means of an online survey (N = 835), we investigated how data subjects expect to be informed about the automated processing of their data. We then conducted a content analysis of the AI disclosures of N = 100 companies and organizations. The combined findings reveal that current GDPR-mandated disclosures do not meet the expectations and needs of data subjects. Explanations drawn up following the guidelines of the generic formulations of the GDPR differ widely and are often vague, incomplete and lack transparency. In our conclusions we identify a path towards standardizing and optimizing AI information notices. (shrink)

This paper considers challenges resulting from the use of advanced artificial judicial intelligence (AAJI). We argue that these challenges should be considered through the lens of value alignment. Instead of discussing why specific goals and values, such as fairness and nondiscrimination, ought to be implemented, we consider the question of how AAJI can be aligned with goals and values more generally, in order to be reliably integrated into legal and judicial systems. This value alignment framing draws on AI safety and (...) alignment literature to introduce two otherwise neglected considerations for AAJI safety: specification and assurance. We outline diverse research directions and suggest the adoption of assurance and specification mechanisms as the use of AI in the judiciary progresses. While we focus on specification and assurance to illustrate the value of the AI safety and alignment literature, we encourage researchers in law and philosophy to consider what other lessons may be drawn. (shrink)

We propose a formal framework for interpretable machine learning. Combining elements from statistical learning, causal interventionism, and decision theory, we design an idealised explanation game in which players collaborate to find the best explanation for a given algorithmic prediction. Through an iterative procedure of questions and answers, the players establish a three-dimensional Pareto frontier that describes the optimal trade-offs between explanatory accuracy, simplicity, and relevance. Multiple rounds are played at different levels of abstraction, allowing the players to explore overlapping causal (...) patterns of variable granularity and scope. We characterise the conditions under which such a game is almost surely guaranteed to converge on a optimal explanation surface in polynomial time, and highlight obstacles that will tend to prevent the players from advancing beyond certain explanatory thresholds. The game serves a descriptive and a normative function, establishing a conceptual space in which to analyse and compare existing proposals, as well as design new and improved solutions. (shrink)

We propose a formal framework for interpretable machine learning. Combining elements from statistical learning, causal interventionism, and decision theory, we design an idealisedexplanation gamein which players collaborate to find the best explanation(s) for a given algorithmic prediction. Through an iterative procedure of questions and answers, the players establish a three-dimensional Pareto frontier that describes the optimal trade-offs between explanatory accuracy, simplicity, and relevance. Multiple rounds are played at different levels of abstraction, allowing the players to explore overlapping causal patterns of (...) variable granularity and scope. We characterise the conditions under which such a game is almost surely guaranteed to converge on a (conditionally) optimal explanation surface in polynomial time, and highlight obstacles that will tend to prevent the players from advancing beyond certain explanatory thresholds. The game serves a descriptive and a normative function, establishing a conceptual space in which to analyse and compare existing proposals, as well as design new and improved solutions. (shrink)

As machine learning has gradually entered into ever more sectors of public and private life, there has been a growing demand for algorithmic explainability. How can we make the predictions of complex statistical models more intelligible to end users? A subdiscipline of computer science known as interpretable machine learning (IML) has emerged to address this urgent question. Numerous influential methods have been proposed, from local linear approximations to rule lists and counterfactuals. In this article, I highlight three conceptual challenges that (...) are largely overlooked by authors in this area. I argue that the vast majority of IML algorithms are plagued by (1) ambiguity with respect to their true target; (2) a disregard for error rates and severe testing; and (3) an emphasis on product over process. Each point is developed at length, drawing on relevant debates in epistemology and philosophy of science. Examples and counterexamples from IML are considered, demonstrating how failure to acknowledge these problems can result in counterintuitive and potentially misleading explanations. Without greater care for the conceptual foundations of IML, future work in this area is doomed to repeat the same mistakes. (shrink)

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

Recent algorithmic technologies have challenged law’s anthropocentric assumptions. In this article, we develop a set of theoretical tools drawn from new materialisms and the philosophy of information to unravel the complex intra-actions between law and computer code. Accordingly, we first propose a framework for understanding the enmeshing of law and code based on a diffractive reading of Barad’s agential realism and Simondon’s theory of information. We argue that once law and code are understood as material entities that intra-act through in-formation, (...) the concept of transduction allows us to trace how they push each other towards change. After developing the theoretical tools, we deploy them to make sense of how law and code have changed in response to increasing automation of decision-making and the appearance of unexplainable artificial intelligence (AI) code. Thus, we employ a case study to trace transformations of the right to explanation under the European data protection regulations. This provides the backdrop for our account of how law transduces into code (and vice versa) and a proving ground for our framework. (shrink)

Transparency is now a fundamental principle for data processing under the General Data Protection Regulation. We explore what this requirement entails for artificial intelligence and automated decision-making systems. We address the topic of transparency in artificial intelligence by integrating legal, social, and ethical aspects. We first investigate the ratio legis of the transparency requirement in the General Data Protection Regulation and its ethical underpinnings, showing its focus on the provision of information and explanation. We then discuss the pitfalls with respect (...) to this requirement by focusing on the significance of contextual and performative factors in the implementation of transparency. We show that human–computer interaction and human-robot interaction literature do not provide clear results with respect to the benefits of transparency for users of artificial intelligence technologies due to the impact of a wide range of contextual factors, including performative aspects. We conclude by integrating the information- and explanation-based approach to transparency with the critical contextual approach, proposing that transparency as required by the General Data Protection Regulation in itself may be insufficient to achieve the positive goals associated with transparency. Instead, we propose to understand transparency relationally, where information provision is conceptualized as communication between technology providers and users, and where assessments of trustworthiness based on contextual factors mediate the value of transparency communications. This relational concept of transparency points to future research directions for the study of transparency in artificial intelligence systems and should be taken into account in policymaking. (shrink)

Certain characteristics make machine learning a powerful tool for processing large amounts of data, and also particularly unsuitable for explanatory purposes. There are worries that its increasing use in science may sideline the explanatory goals of research. We analyze the key characteristics of ML that might have implications for the future directions in scientific research: epistemic opacity and the ‘theory-agnostic’ modeling. These characteristics are further analyzed in a comparison of ML with the traditional statistical methods, in order to demonstrate what (...) it is specifically that makes ML methodology substantially unsuitable for reaching explanations. The analysis is given broader philosophical context by connecting it with the views on the role of prediction and explanation in science, their relationship, and the value of explanation. We proceed to show, first, that ML disrupts the relationship between prediction and explanation commonly understood as a functional relationship. Then we show that the value of explanation is not exhausted in purely predictive functions, but rather has a ubiquitously recognized value for both science and everyday life. We then invoke two hypothetical scenarios with different degrees of automatization of science, which help test our intuitions on the role of explanation in science. The main question we address is whether ML will reorient or otherwise impact our standard explanatory practice. We conclude with a prognosis that ML would diversify science into purely predictively oriented research based on ML-like techniques and, on the other hand, remaining faithful to anthropocentric research focused on the search for explanation. (shrink)

The paper has two goals. The first is presenting the main results of the recent report Ethics of Connected and Automated Vehicles: recommendations on road safety, privacy, fairness, explainability and responsibility written by the Horizon 2020 European Commission Expert Group to advise on specific ethical issues raised by driverless mobility, of which the author of this paper has been member and rapporteur. The second is presenting some broader ethical and philosophical implications of these recommendations, and using these to contribute to (...) the establishment of Ethics of Transportation as an independent branch of applied ethics. The recent debate on the ethics of Connected and Automated Vehicles presents a paradox and an opportunity. The paradox is the presence of a flourishing debate on the ethics of one very specific transportation technology without ethics of transportation being in itself a well-established academic discipline. The opportunity is that now that a spotlight has been switched on the ethical dimensions of CAVs it may be easier to establish a broader debate on ethics of transportation. While the 20 recommendations of the EU report are grouped in three macro-areas: road safety, data ethics, and responsibility, in this paper they will be grouped according to eight philosophical themes: Responsible Innovation, road justice, road safety, freedom, human control, privacy, data fairness, responsibility. These are proposed as the first topics for a new ethics of transportation. (shrink)

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

When agents insert technological systems into their decision-making processes, they can obscure moral responsibility for the results. This can give rise to a distinct moral wrong, which we call “agency laundering.” At root, agency laundering involves obfuscating one’s moral responsibility by enlisting a technology or process to take some action and letting it forestall others from demanding an account for bad outcomes that result. We argue that the concept of agency laundering helps in understanding important moral problems in a number (...) of recent cases involving automated, or algorithmic, decision-systems. We apply our conception of agency laundering to a series of examples, including Facebook’s automated advertising suggestions, Uber’s driver interfaces, algorithmic evaluation of K-12 teachers, and risk assessment in criminal sentencing. We distinguish agency laundering from several other critiques of information technology, including the so-called “responsibility gap,” “bias laundering,” and masking. (shrink)

There is widespread agreement that there should be a principle requiring that artificial intelligence be ‘explicable’. Microsoft, Google, the World Economic Forum, the draft AI ethics guidelines for the EU commission, etc. all include a principle for AI that falls under the umbrella of ‘explicability’. Roughly, the principle states that “for AI to promote and not constrain human autonomy, our ‘decision about who should decide’ must be informed by knowledge of how AI would act instead of us” :689–707, 2018). There (...) is a strong intuition that if an algorithm decides, for example, whether to give someone a loan, then that algorithm should be explicable. I argue here, however, that such a principle is misdirected. The property of requiring explicability should attach to a particular action or decision rather than the entity making that decision. It is the context and the potential harm resulting from decisions that drive the moral need for explicability—not the process by which decisions are reached. Related to this is the fact that AI is used for many low-risk purposes for which it would be unnecessary to require that it be explicable. A principle requiring explicability would prevent us from reaping the benefits of AI used in these situations. Finally, the explanations given by explicable AI are only fruitful if we already know which considerations are acceptable for the decision at hand. If we already have these considerations, then there is no need to use contemporary AI algorithms because standard automation would be available. In other words, a principle of explicability for AI makes the use of AI redundant. (shrink)

With Artificial Intelligence entering our lives in novel ways—both known and unknown to us—there is both the enhancement of existing ethical issues associated with AI as well as the rise of new ethical issues. There is much focus on opening up the ‘black box’ of modern machine-learning algorithms to understand the reasoning behind their decisions—especially morally salient decisions. However, some applications of AI which are no doubt beneficial to society rely upon these black boxes. Rather than requiring algorithms to be (...) transparent we should focus on constraining AI and those machines powered by AI within microenvironments—both physical and virtual—which allow these machines to realize their function whilst preventing harm to humans. In the field of robotics this is called ‘envelopment’. However, to put an ‘envelope’ around AI-powered machines we need to know some basic things about them which we are often in the dark about. The properties we need to know are the: training data, inputs, functions, outputs, and boundaries. This knowledge is a necessary first step towards the envelopment of AI-powered machines. It is only with this knowledge that we can responsibly regulate, use, and live in a world populated by these machines. (shrink)

With the increasing prevalence of artificial intelligence (AI) and other digital technologies in healthcare, the ethical debate surrounding their adoption is becoming more prominent. Here I consider the issue of gaining informed patient consent to AI-enhanced care from the vantage point of the United Kingdom’s National Health Service setting. I build my discussion around two claims from the World Health Organization: that healthcare services should not be denied to individuals who refuse AI-enhanced care and that there is no precedence to (...) seeking patient consent to AI-enhanced care. I discus U.K. law relating to patient consent and the General Data Protection Regulation to show that current standards relating to patient consent are adequate for AI-enhanced care. I then suggest that in the future it may not be possible to guarantee patient access to non-AI-enhanced healthcare, in a similar way to how we do not offer patients manual alternatives to automated healthcare processes. Throughout my discussion I focus on the issues of patient choice and veracity in the patient–clinician relationship. Finally, I suggest that the best way to protect patients from potential harms associated with the introduction of AI to patient care is not via an overly burdensome patient consent process but via evaluation and regulation of AI technologies. (shrink)

Assistive systems based on Artificial Intelligence (AI) are bound to reshape decision-making in all areas of society. One of the most intricate challenges arising from their implementation in high-stakes environments such as medicine concerns their frequently unsatisfying levels of explainability, especially in the guise of the so-called black-box problem: highly successful models based on deep learning seem to be inherently opaque, resisting comprehensive explanations. This may explain why some scholars claim that research should focus on rendering AI systems understandable, rather (...) than explainable. Yet, there is a grave lack of agreement concerning these terms in much of the literature on AI. We argue that the seminal distinction made by the philosopher and physician Karl Jaspers between different types of explaining and understanding in psychopathology can be used to promote greater conceptual clarity in the context of Machine Learning (ML). Following Jaspers, we claim that explaining and understanding constitute multi-faceted epistemic approaches that should not be seen as mutually exclusive, but rather as complementary ones as in and of themselves they are necessarily limited. Drawing on the famous example of Watson for Oncology we highlight how Jaspers’ methodology translates to the case of medical AI. Classical considerations from the philosophy of psychiatry can therefore inform a debate at the centre of current AI ethics, which in turn may be crucial for a successful implementation of ethically and legally sound AI in medicine. (shrink)

The paper focuses on concerns and legal challenges brought on by the use of algorithms. A particular class of algorithms that augment or replace analysis and decision-making by humans, i.e. data analytics and machine learning, is under scrutiny. Taking into account Balkin’s work on “the laws of an algorithmic society”, attention is drawn to obligations of transparency, matters of due process, and accountability. This US-centric analysis on drawbacks and loopholes of current legal systems is complemented with the analysis of norms (...) and principles of the EU data protection law, or “GDPR”. The aim is twofold. On the one hand, the intent is to shed light on some crucial differences between the US and EU law on the regulation of algorithmic operators, both public and private. Whereas, in the USA, scholars debate whether and to what extent new duties and responsibilities of algorithmic operators, e.g. information fiduciaries, have to amend the current framework of self-regulation and light government—as shown by the White House’s Office of Science and Technology Policy report from November 2016—in EU law much of the new duties and responsibilities of algorithmic operators have been passed upon them as data controllers. Whether such approaches will successfully tackle the normative challenges of the algorithmic society is, on the other hand, an open issue that will likely represent the main topic of debate over the next years. Disagreement may concern: the terms framing the legal question in e.g. statistical purposes of the data processing; how such terms are related to each other in legal reasoning ; and legal hard cases that will increasingly have to do with the principles that are at stake also but not only in data protection. By entrusting such legal hard cases to algorithms, or some sort of smart artificial agent, humans still bear full responsibility for the judgment of what is socially, ethically, and legally “plain” and “hard” in social affairs. The balance between delegation of decisions to algorithms and non-delegation will be the leitmotiv of the algorithmic society. Since the devil is in the detail, the current paper is devoted to some of them. (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)

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 potential for biases being built into algorithms has been known for some time, yet literature has only recently demonstrated the ways algorithmic profiling can result in social sorting and harm marginalised groups. We contend that with increased algorithmic complexity, biases will become more sophisticated and difficult to identify, control for, or contest. Our argument has four steps: first, we show how harnessing algorithms means that data gathered at a particular place and time relating to specific persons, can be used (...) to build group models applied in different contexts to different persons. Thus, privacy and data protection rights, with their focus on individuals, do not protect from the discriminatory potential of algorithmic profiling. Second, we explore the idea that anti-discrimination regulation may be more promising, but acknowledge limitations. Third, we argue that in order to harness anti-discrimination regulation, it needs to confront emergent forms of discrimination or risk creating new invisibilities, including invisibility from existing safeguards. Finally, we outline suggestions to address emergent forms of discrimination and exclusionary invisibilities via intersectional and post-colonial analysis. (shrink)

The increasing use of Artificial Intelligence for making decisions in public affairs has sparked a lively debate on the benefits and potential harms of self-learning technologies, ranging from the hopes of fully informed and objectively taken decisions to fear for the destruction of mankind. To prevent the negative outcomes and to achieve accountable systems, many have argued that we need to open up the “black box” of AI decision-making and make it more transparent. Whereas this debate has primarily focused on (...) how transparency can secure high-quality, fair, and reliable decisions, far less attention has been devoted to the role of transparency when it comes to how the general public come to perceive AI decision-making as legitimate and worthy of acceptance. Since relying on coercion is not only normatively problematic but also costly and highly inefficient, perceived legitimacy is fundamental to the democratic system. This paper discusses how transparency in and about AI decision-making can affect the public’s perception of the legitimacy of decisions and decision-makers and produce a framework for analyzing these questions. We argue that a limited form of transparency that focuses on providing justifications for decisions has the potential to provide sufficient ground for perceived legitimacy without producing the harms full transparency would bring. (shrink)

“Rationality” in Simon's “bounded rationality” is the principle that humans make decisions on the basis of step-by-step reasoning using systematic rules of logic to maximize utility. “Bounded rationality” is the observation that the ability of a human brain to handle algorithmic complexity and large quantities of data is limited. Bounded rationality, in other words, treats a decision maker as a machine carrying out computations with limited resources. Under the principle of embodied cognition, a cognitive mind is an interactive machine. Turing-Church (...) computations are not interactive, and interactive machines can accomplish things that no Turing-Church computation can accomplish. Hence, if “rationality” is computation, and “bounded rationality” is computation with limited complexity, then “embodied bounded rationality” is both more limited than computation and more powerful. By embracing interaction, embodied bounded rationality can accomplish things that Turing-Church computation alone cannot. Deep neural networks, which have led to a revolution in artificial intelligence, are both interactive and not fundamentally algorithmic. Hence, their ability to mimic some cognitive capabilities far better than prior algorithmic techniques based on symbol manipulation provides empirical evidence for the principle of embodied bounded rationality. (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)

A growing literature is taking an institutionalist and governance perspective on how algorithms shape society based on unprecedented capacities for managing social complexity. Algorithmic governance altogether emerges as a novel and distinctive kind of societal steering. It appears to transcend established categories and modes of governance—and thus seems to call for new ways of thinking about how social relations can be regulated and ordered. However, as this paper argues, despite its novel way of realizing outcomes of collective steering and coordination, (...) it can nevertheless be grasped with an old and fundamental figure in political philosophy: that of Thomas Hobbes’ Leviathan. Comparing algorithmic governance with this figure serves to highlight their similarities as socio-political arrangements, and specifically to clarify how algorithmic governance parallels the apolitical traits of the Leviathan—it eliminates the political as it requires compliance and forgoing contestation to best fulfill its role and to produce satisfying outcomes. (shrink)

This essay provides a framework of concepts and principles suitable for systematic discussion of issues surrounding expertise. Expertise creates inequality. Its multiple benefits and the creativity of technology lead to a society replete with expertises. The basic binds of expertise derive from the desire of non-experts to be able to both enjoy what expertise offers and insure that it is exercised in the social interest. This involves trusting the exercise of expertise, involuntarily or voluntarily. A healthy society provides various means (...) to move trust from involuntary to voluntary. The social means for achieving this are laid out. The purpose of this short essay is to briefly lay out a conceptual framework within which to construct, clarify, evaluate and apply expertises. It is not to promote some particular notion of expertise over others, or to review the vast literatures, such as that on trust in science, that make up the domain. A few notes on one work towards this essay’s close may indicate what a major, and expert, process this would be. (shrink)

Rapid developments in Artificial Intelligence are leading to an increasing human reliance on machine decision making. Even in collaborative efforts with Decision Support Systems (DSSs), where a human expert is expected to make the final decisions, it can be hard to keep the expert actively involved throughout the decision process. DSSs suggest their own solutions and thus invite passive decision making. To keep humans actively ‘on’ the decision-making loop and counter overreliance on machines, we propose a ‘reflection machine’ (RM). This (...) system asks users questions about their decision strategy and thereby prompts them to evaluate their own decisions critically. We discuss what forms RMs can take and present a proof-of-concept implementation of a RM that can produce feedback on users’ decisions in the medical and law domains. We show that the prototype requires very little domain knowledge to create reasonably intelligent critiquing questions. With this prototype, we demonstrate the technical feasibility to develop RMs and hope to pave the way for future research into their effectiveness and value. (shrink)

Human decisions are increasingly supported by decision support systems (DSS). Humans are required to remain “on the loop,” by monitoring and approving/rejecting machine recommendations. However, use of DSS can lead to overreliance on machines, reducing human oversight. This paper proposes “reflection machines” (RM) to increase meaningful human control. An RM provides a medical expert not with suggestions for a decision, but with questions that stimulate reflection about decisions. It can refer to data points or suggest counterarguments that are less compatible (...) with the planned decision. RMs think against the proposed decision in order to increase human resistance against automation complacency. Building on preliminary research, this paper will (1) make a case for deriving a set of design requirements for RMs from EU regulations, (2) suggest a way how RMs could support decision-making, (3) describe the possibility of how a prototype of an RM could apply to the medical domain of chronic low back pain, and (4) highlight the importance of exploring an RM’s functionality and the experiences of users working with it. (shrink)

AI-based algorithms are used extensively by public institutions. Thus, for instance, AI algorithms have been used in making decisions concerning punishment providing welfare payments, making decisions concerning parole, and many other tasks which have traditionally been assigned to public officials and/or public entities. We develop a novel argument against the use of AI algorithms, in particular with respect to decisions made by public officials and public entities. We argue that decisions made by AI algorithms cannot count as public decisions, namely (...) decisions that are made in the name of citizens and that this fact should be taken into consideration when utilizing AI to replace public officials. (shrink)

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)

The field of computational law has increasingly moved into the focus of the scientific community, with recent research analysing its issues and risks. In this article, we seek to draw a structured and comprehensive list of societal issues that the deployment of automatically processable regulation could entail. We do this by systematically exploring attributes of the law that are being challenged through its encoding and by taking stock of what issues current projects in this field raise. This article adds to (...) the current literature not only by providing a needed framework to structure arising issues of computational law but also by bridging the gap between theoretical literature and practical implementation. Key findings of this article are: (1) The primary benefit (efficiency vs. accessibility) sought after when encoding law matters with respect to the issues such an endeavor triggers; (2) Specific characteristics of a project—project type, degree of mediation by computers, and potential for divergence of interests—each impact the overall number of societal issues arising from the implementation of automatically processable regulation. (shrink)

The United Nations confirmed that privacy remains a human right in the digital age, but our daily digital experiences and seemingly ever-increasing amounts of data suggest that privacy is a mundane, distributed and technologically mediated concept. This article explores privacy by mapping out different legal and conceptual approaches to privacy protection in the context of datafication. It provides an essential starting point to explore the entwinement of technological, ethical and regulatory dynamics. It clarifies why each of the presented approaches emphasises (...) particular aspects and analyses the tensions that arise. The resulting overview provides insight into the main strengths and limitations of the different approaches arising from specific traditions. This analytic overview therefore serves as a key resource to analyse the usefulness of the approaches in the context of the increasing datafication of both private and public spheres.Specifically, we contrast the approach focusing on data subjects whose data are being ‘protected’ with others, including Fair Information Practice Principles, the German right to ‘informational self-determination’, and the South American ‘habeas data’ doctrine. We also present and contrast emerging approaches to privacy and discuss their intersection with datafication. In conclusion, we put forth that rather than aiming for one single solution that works worldwide and across all situations, it is essential to identify synergies and stumbling blocks between the various regulatory settings and newly emerging approaches. (shrink)

In light of recent advances in machine learning for medical applications, the automation of medical diagnostics is imminent. That said, before machine learning algorithms find their way into clinical practice, various problems at the epistemic level need to be overcome. In this paper, we discuss different sources of uncertainty arising for clinicians trying to evaluate the trustworthiness of algorithmic evidence when making diagnostic judgments. Thereby, we examine many of the limitations of current machine learning algorithms (with deep learning in particular) (...) and highlight their relevance for medical diagnostics. Among the problems we inspect are the theoretical foundations of deep learning (which are not yet adequately understood), the opacity of algorithmic decisions, and the vulnerabilities of machine learning models, as well as concerns regarding the quality of medical data used to train the models. Building on this, we discuss different desiderata for an uncertainty amelioration strategy that ensures that the integration of machine learning into clinical settings proves to be medically beneficial in a meaningful way. (shrink)

The idea of artificial intelligence for social good is gaining traction within information societies in general and the AI community in particular. It has the potential to tackle social problems through the development of AI-based solutions. Yet, to date, there is only limited understanding of what makes AI socially good in theory, what counts as AI4SG in practice, and how to reproduce its initial successes in terms of policies. This article addresses this gap by identifying seven ethical factors that are (...) essential for future AI4SG initiatives. The analysis is supported by 27 case examples of AI4SG projects. Some of these factors are almost entirely novel to AI, while the significance of other factors is heightened by the use of AI. From each of these factors, corresponding best practices are formulated which, subject to context and balance, may serve as preliminary guidelines to ensure that well-designed AI is more likely to serve the social good. (shrink)

Real engines of the artificial intelligence revolution, machine learning models, and algorithms are embedded nowadays in many services and products around us. As a society, we argue it is now necessary to transition into a phronetic paradigm focused on the ethical dilemmas stemming from the conception and application of AIs to define actionable recommendations as well as normative solutions. However, both academic research and society-driven initiatives are still quite far from clearly defining a solid program of study and intervention. In (...) this contribution, we will focus on selected ethical investigations around AI by proposing an incremental model of trust that can be applied to both human-human and human-AI interactions. Starting with a quick overview of the existing accounts of trust, with special attention to Taddeo’s concept of “e-trust,” we will discuss all the components of the proposed model and the reasons to trust in human-AI interactions in an example of relevance for business organizations. We end this contribution with an analysis of the epistemic and pragmatic reasons of trust in human-AI interactions and with a discussion of kinds of normativity in trustworthiness of AIs. (shrink)

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

The creation of increasingly complex artificial intelligence (AI) systems raises urgent questions about their ethical and social impact on society. Since this impact ultimately depends on political decisions about normative issues, political philosophers can make valuable contributions by addressing such questions. Currently, AI development and application are to a large extent regulated through non-binding ethics guidelines penned by transnational entities. Assuming that the global governance of AI should be at least minimally democratic and fair, this paper sets out three desiderata (...) that an account should satisfy when theorizing about what this means. We argue, first, that an analysis of democratic values, political entities and decision-making should be done in a holistic way; second, that fairness is not only about how AI systems treat individuals, but also about how the benefits and burdens of transformative AI are distributed; and finally, that justice requires that governance mechanisms are not limited to AI technology, but are incorporated into a range of basic institutions. Thus, rather than offering a substantive theory of democratic and fair AI governance, our contribution is metatheoretical: we propose a theoretical framework that sets up certain normative boundary conditions for a satisfactory account. (shrink)

Should we be allowed to refuse any involvement of artificial intelligence technology in diagnosis and treatment planning? This is the relevant question posed by Ploug and Holm in a recent article in Medicine, Health Care and Philosophy. In this article, I adhere to their conclusions, but not necessarily to the rationale that supports them. First, I argue that the idea that we should recognize this right on the basis of a rational interest defence is not plausible, unless we are willing (...) to judge each patient’s ideology or religion. Instead, I consider that the right must be recognized by virtue of values such as social pluralism or individual autonomy. Second, I point out that the scope of such a right should be limited at least under three circumstances: if it is against a physician’s obligation to not cause unnecessary harm to a patient or to not provide futile treatment, in cases where the costs of implementing this right are too high, or if recognizing the right would deprive other patients of their own rights to adequate health care. (shrink)

Algorithmic decision-making based on profiling may significantly affect people’s destinies. As a rule, however, explanations for such decisions are lacking. What are the chances for a “right to explanation” to be realized soon? After an exploration of the regulatory efforts that are currently pushing for such a right it is concluded that, at the moment, the GDPR stands out as the main force to be reckoned with. In cases of profiling, data subjects are granted the right to receive meaningful information (...) about the functionality of the system in use; for fully automated profiling decisions even an explanation has to be given. However, the trade secrets and intellectual property rights (IPRs) involved must be respected as well. These conflicting rights must be balanced against each other; what will be the outcome? Looking back to 1995, when a similar kind of balancing had been decreed in Europe concerning the right of access (DPD), Wachter et al. (2017) find that according to judicial opinion only generalities of the algorithm had to be disclosed, not specific details. This hardly augurs well for a future right of access let alone to explanation. Thereupon the landscape of IPRs for machine learning (ML) is analysed. Spurred by new USPTO guidelines that clarify when inventions are eligible to be patented, the number of patent applications in the US related to ML in general, and to “predictive analytics” in particular, has soared since 2010—and Europe has followed. I conjecture that in such a climate of intensified protection of intellectual property, companies may legitimately claim that the more their application combines several ML assets that, in addition, are useful in multiple sectors, the more value is at stake when confronted with a call for explanation by data subjects. Consequently, the right to explanation may be severely crippled. (shrink)

There is general consensus that explainable artificial intelligence is valuable, but there is significant divergence when we try to articulate why, exactly, it is desirable. This question must be distinguished from two other kinds of questions asked in the XAI literature that are sometimes asked and addressed simultaneously. The first and most obvious is the ‘how’ question—some version of: ‘how do we develop technical strategies to achieve XAI?’ Another question is specifying what kind of explanation is worth having in the (...) first place. As difficult and important as the challenges are in answering these questions, they are distinct from a third question: why do we want XAI at all? There is vast literature on this question as well, but I wish to explore a different kind of answer. The most obvious way to answer this question is by describing a desirable outcome that would likely be achieved with the right kind of explanation, which would make the explanation valuable instrumentally. That is, XAI is desirable to attain some other value, such as fairness, trust, accountability, or governance. This family of arguments is obviously important, but I argue that explanations are also intrinsically valuable, because unexplainable systems can be dehumanizing. I argue that there are at least three independently valid versions of this kind of argument: an argument from participation, from knowledge, and from actualization. Each of these arguments that XAI is intrinsically valuable is independently compelling, in addition to the more obvious instrumental benefits of XAI. (shrink)

Over the last decade, humans have produced each year as much data as were produced throughout the entire history of humankind. These data, in quantities that exceed current analytical capabilities, have been described as “the new oil,” an incomparable source of value. This is true for healthcare, as well. Conducting analyses of large, diverse, medical datasets promises the detection of previously unnoticed clinical correlations and new diagnostic or even therapeutic possibilities. However, using Big Data poses several problems, especially in terms (...) of representing the uniqueness of each patient and expressing the differences between individuals, primarily gender and sex differences. The first two sections of the paper provide a definition of “Big Data” and illustrate the uses of Big Data in medicine. Subsequently, the paper explores the struggle to represent exhaustively the uniqueness of the patient through Big Data is highlighted prior to a deeper investigation of the digital representation of gender in personalized medicine. The final part of the paper put forward a series of recommendations for better approaching the complexity of gender in medical and clinical research involving Big Data for the creation or enhancement of personalized medicine services. (shrink)

Advancements in machine learning have fuelled the popularity of using AI decision algorithms in procedures such as bail hearings, medical diagnoses and recruitment. Academic articles, policy texts, and popularizing books alike warn that such algorithms tend to be opaque: they do not provide explanations for their outcomes. Building on a causal account of transparency and opacity as well as recent work on the value of causal explanation, I formulate a moral concern for opaque algorithms that is yet to receive a (...) systematic treatment in the literature: when such algorithms are used in life-changing decisions, they can obstruct us from effectively shaping our lives according to our goals and preferences, thus undermining our autonomy. I argue that this concern deserves closer attention as it furnishes the call for transparency in algorithmic decision-making with both new tools and new challenges. (shrink)