Ethics of Artificial Intelligence

On the possibilities and limits of the ethics of artificial intelligence. An overview of current developments and debates with a focus on language processing systems. -/- Driven by the success of artificial intelligence (AI), the ethics of AI is currently enjoying a boom. Advice from ethics experts is increasingly being sought by policymakers and industry to proactively identify the risks associated with new AI technologies and to propose solutions. But how realistic are the expectations placed on AI ethics to make (...) technological development more equitable and acceptable? Doesn't AI ethics itself run the risk of becoming a fig leaf in the service of pre-determined economic and political goals? Based on these questions, this article presents and discusses some of the most current debates and proposals for dealing with the opportunities and risks of AI technologies, with a particular focus on language technologies. Although the danger of instrumentalization is present, the article concludes, the ethics of AI currently represents an important opportunity to identify problematic developments at an early stage and to initiate societal debates, political processes, and legal regulations that can make future AI developments more equitable and environmentally sound. (shrink)

Artificial intelligence (AI) and molecular biotechnologies (MB) are among the most promising, but also ethically hotly debated emerging technologies. In both fields, several ethics reports, which invoke lists of ethics principles, have been put forward. These reports and the principles lists are technology specific. This article aims to contribute to the ongoing debate on ethics of emerging technologies by comparatively analysing four European ethics reports from the two technology fields. Adopting a qualitative and in-depth approach, the article highlights how ethics (...) principles from MB can inform AI ethics and vice versa. By synthesizing the respective ethical cores of the principles included in the analysed reports, the article derives, moreover, a unified list of principles for assessing emerging technologies. The suggested list consists of nine principles: autonomy; individual and social well-being and prevention of harm; reliability, safety and security; informational privacy; transparency; accountability; communication, participation and democracy; justice, fairness, and non-discrimination; sustainability. (shrink)

1. I am, in fact, a person – Moralischer Status von KI 2. Superintelligenz – Ende oder Rettung der Menschheit? 3. Diskriminierung – KI als Ursache oder Lösung? -/- Authors: Dominik Balg, Larissa Bolte, Anne Burkard, Jan Constantin, Leonard Dung, Jürn Gottschalk, Kerstin Gregor-Gehrmann, Isabelle Guntermann und Katharina Schulz.

This paper aims to engineer the concept of biomimetic design for its application in agricultural technology as an innovation strategy to sustain non-human species’ adaptation to today’s rapid environmental changes. By questioning the alleged intrinsic morality of biomimicry, a formulation of it is sought that goes beyond the sharp distinction between nature as inspiration and the human field of application of biomimetic technologies. After reviewing the main literature on Responsible Innovation, we support Vincent Blok’s “eco-centric” perspective on biomimicry, which considers (...) the human and natural worlds as indistinguishable parts of a shared Earth. We propose this approach as a complement to the “evomimetic” critique, which warns biomimetic research against the limits of adaptationism. By merging these two reframing of the biomimicry concept, we thus pave the way for a new understanding of the use of human-inspired technology (such as artificial intelligence) to help the “evolution” of domesticated species into (semi)autonomous natural-technological hybrids. In particular, the examples we consider concern the potential of AI-enabled robotic bases in technological beekeeping. (shrink)

Attempts to influence people’s beliefs through misinformation have a long history. In the age of social media, however, there is a growing fear that the circulation of false or misleading claims will be more impactful than ever now that sophisticated technological means are available to those who desire to spread them. Should democratic societies worry about misinformation? If so, is it possible and desirable for them to control its spread by regulating it? This chapter offers an answer to these questions. (...) First, I propose a definition of misinformation and explain how it proliferates in online contexts. Second, I consider four reasons to worry about misinformation by discussing its likely impact on people’s political opinions, emotions, physical safety and personal autonomy. Third, I assess three strategies for regulating misinformation—individual self-regulation, platform-based innovations and governmental action—and suggest that the most effective ones are those that spare human agents from having to successively review individual pieces of content. (shrink)

This paper aims to clear up the epistemology of learning morality from Artificial Moral Advisors (AMAs). We start with a brief consideration of what counts as moral enhancement and consider the risk of deskilling raised by machines that offer moral advice. We then shift focus to the epistemology of moral advice and show when and under what conditions moral advice can lead to enhancement. We argue that people’s motivational dispositions are enhanced by inspiring people to act morally, instead of merely (...) telling them how to act. Drawing upon these insights, we claim that if AMAs are to genuinely enhance people morally, they should be designed as inspiration and not authority machines. In the final section, we evaluate existing AMA models to shed light on which holds the most promise for helping to make users better moral agents. (shrink)

Applications and devices of artificial intelligence are increasingly common in the healthcare field. Robots fulfilling some caregiving functions are not a distant future. In this scenario, we must ask ourselves if it is possible for machines to care to the extent of completely replacing human care and if such replacement, if possible, is desirable. In this paper, I argue that caregiving requires know-how permeated by affectivity that is far from being achieved by currently available machines. I also maintain that the (...) complete substitution of machine caregiving for human caregiving is not desirable because caregiving requires genuine human connection. (shrink)

Virtue theory is a natural approach towards the design of artificially intelligent systems, given that the design of artificial intelligence essentially aims at designing agents with excellent dispositions. This has led to a lively research programme to develop artificial virtues. However, this research programme has until now had a narrow focus on moral virtues in an Aristotelian mould. While Aristotelian moral virtue has played a foundational role for the field, it unduly constrains the possibilities of virtue theory for artificial intelligence. (...) This paper aims to remedy this limitation. Philosophers have developed a rich tradition investigating virtues, their normative domains and their structure. Drawing on this tradition, I propose a three-dimensional classification system of possible artificial virtues: virtues can be classified according to the domain in which virtue is an excellence, norm which makes a virtue an excellence, and mode how the virtue delivers the excellence. With this framework, we can discern gaps in the current theorising about artificial virtues. Additionally, it gives us a tool to evaluate the competences of extant artificially intelligent systems. (shrink)

In this paper, I argue that surveillance capitalism and computational propaganda can undermine democratic equality. First, I argue that two types of resources are relevant for democratic equality: 1) free time, which entails time that is free from systemic surveillance, and 2) epistemic resources. In order for everyone in a democratic system to be equally capable of full political participation, it’s a minimum requirement that these two resources are distributed fairly. But AI that’s used for surveillance capitalism can undermine the (...) fair distribution of these resources, thereby threatening democracy. I further argue that computational propaganda undermines the democratic aim of collective self-determination by normalizing relations of domination and thereby disrupting the equal standing of persons. I conclude by considering some potential solutions. (shrink)

Artificial intelligent technologies are utilized to provide online personalized recommendations, suggestions, or prompts that can influence people's decision-making processes. We call this AI-enhanced nudging (or AI-nudging for short). Contrary to the received wisdom we claim that AI-enhanced nudging is not necessarily morally problematic. To start assessing the risks and moral import of AI-nudging we believe that we should adopt a risk-factor analysis: we show that both the level of risk and possibly the moral value of adopting AI-nudging ultimately depend on (...) the complexity of the “moral situation” in which various risk-factors play a causal role in determining levels of uncertainty. (shrink)

The rapid integration of these artificial intelligence (AI) tools into our daily lives is reshaping how we think and make decisions. We propose that data-driven AI systems, by transcending individual artefacts and interfacing with a dynamic, multiartefact ecosystem, constitute a distinct psychological system. We call this ‘system 0’, and position it alongside Kahneman’s system 1 (fast, intuitive thinking) and system 2 (slow, analytical thinking).System 0 represents the outsourcing of certain cognitive tasks to AI, which can process vast amounts of data (...) and perform complex computations beyond human capabilities. It emerges from the interaction between users and AI systems, which creates a dynamic, personalized interface. Authors: Massimo Chiriatti, Marianna Ganapini, Enrico Panai, Mario Ubiali & Giuseppe Riva. (shrink)

This paper argues that the emotional cues exhibited by AI systems designed for social interaction may lead human users to hold misplaced trust in such AI systems, and this poses a substantial problem for human-AI relationships. It begins by discussing the communicative role of certain emotions relevant to perceived trustworthiness. Since displaying such emotions is a reliable indicator of trustworthiness in humans, we use such emotions to assess agents’ trustworthiness according to certain generalizations of folk psychology. Our tendency to engage (...) in anthropomorphism with artificial agents that display these communicative emotions may mislead us into trusting them, even when such trust is misplaced. This phenomenon is referred to as the Emotions-Trustworthiness Gap. This paper demonstrates this general concept with compassion and guilt. In both cases, we may be tempted to judge that AI systems' mere mimicking of relevant emotions constitutes evidence for their trustworthiness. However, most current AI systems are not trustworthy on any major view of trustworthy AI. Thus, the trust elicited from human users will be misplaced. This tendency of human users to hold misplaced trust in their AI companions leaves them vulnerable to serious harm. Several potential solutions to the Emotions-Trustworthiness Gap are discussed. (shrink)

If AI can write an excellent philosophy paper, we argue that philosophy journals should strongly consider publishing that paper. After all, AI stands to make significant contributions to ongoing projects in some subfields, and it benefits the world of philosophy for those contributions to be published in journals, the primary purpose of which is to disseminate significant contributions to philosophy. We also propose the Sponsorship Model of AI journal refereeing to mitigate any costs associated with our view. This model requires (...) at least one human to sponsor any AI manuscript submitted to a journal. After making this case, we evaluate a variety of objections—that our proposal will crash the academic job market, disincentivize humans from doing philosophy, and violate a variety of justified publishing norms—and conclude that they are all wanting. (shrink)

In this essay, I review one of my beloved fictional titles, Wild Wise Weird: The Kingfisher Story collection. The minimal sense of humor and satire in storytelling of Wild Wise Weird are sure to bring readers smiles, better yet, moments of quiet reflection, a much under-appreciated remedy in the world driven almost insane with the abundance of information co-created with AI technologies. I hope to deliver justice to the book.

Natural Language Processing (NLP) research on AI Safety and social bias in AI has focused on safety for humans and social bias against human minorities. However, some AI ethicists have argued that the moral significance of nonhuman animals has been ignored in AI research. Therefore, the purpose of this study is to investigate whether there is speciesism, i.e., discrimination against nonhuman animals, in NLP research. First, we explain why nonhuman animals are relevant in NLP research. Next, we survey the findings (...) of existing research on speciesism in NLP researchers, data, and models and further investigate this problem in this study. The findings of this study suggest that speciesism exists within researchers, data, and models, respectively. Specifically, our survey and experiments show that (a) among NLP researchers, even those who study social bias in AI, do not recognize speciesism or speciesist bias; (b) among NLP data, speciesist bias is inherent in the data annotated in the datasets used to evaluate NLP models; (c) OpenAI GPTs, recent NLP models, exhibit speciesist bias by default. Finally, we discuss how we can reduce speciesism in NLP research. (shrink)

Traditional medical Artificial Intelligence models, approved for clinical use, restrict themselves to single-modal data e.g. images only, limiting their applicability in the complex, multimodal environment of medical diagnosis and treatment. Multimodal Transformer Models in healthcare can effectively process and interpret diverse data forms such as text, images, and structured data. They have demonstrated impressive performance on standard benchmarks like USLME question banks and continue to improve with scale. However, the adoption of these advanced AI models is not without challenges. While (...) multimodal deep learning models like Transformers offer promising advancements in healthcare, their integration requires careful consideration of the accompanying ethical and environmental challenges. (shrink)

There is an underappreciated respect in which the widespread availability of generative artificial intelligence (AI) models poses a threat to human connection. My central contention is that human creativity is especially capable of helping us connect to others in a valuable way, but the widespread availability of generative AI models reduces our incentives to engage in various sorts of creative work in the arts and sciences. I argue that creative endeavors must be motivated by curiosity, and so they must disclose (...) the creative agent’s inquisitive self. It is through self-disclosure, including the disclosure of the inquisitive self, that we put ourselves in a position to be seen by and connect with others through our creative pursuits. Because relying on AI for certain generative tasks is less self-disclosive than the creative work such technologies supplant, this reliance threatens to weaken our connections to one another. (shrink)

Automated influence is one of the most pervasive applications of artificial intelligence in our day-to-day lives, yet a thoroughgoing account of its associated individual and societal harms is lacking. By far the most widespread, compelling, and intuitive account of the harms associated with automated influence follows what I call the control argument. This argument suggests that users are persuaded, manipulated, and influenced by automated influence in a way that they have little or no control over. Based on evidence about the (...) effectiveness of targeted advertising as well as empirical results about the nature of attentional control, I provide reasons to reject this argument. In turn, I use C. Thi Nguyen's theory of value collapse to develop a new account of the harmfulness of automated influence. (shrink)

Critics have raised many moral and legal concerns about posthumous digital avatars. Here my focus instead falls on whether they are likely to enable the bonds with the dead that users apparently yearn for. I conclude that though posthumous avatars can have short-term therapeutic benefits in replicating “habits of intimacy” with the dead, users’ expectations for sustaining long-term bonds with the deceased via posthumous avatars are unlikely to be fulfilled. Posthumous avatars are unlikely to foster the construction of valued memories (...) of the deceased and could in fact impede this process. Furthermore, bonds that develop with posthumous avatars are likely to disappoint users’ expectations for relationships that echo relationships among the living in being temporally dynamic, open-ended, authentic, and mutually shaping. Simply put, users will struggle to sustain living relationships with dead people. Posthumous avatars are therefore likely to prove ill-suited to continue the bonds that the living have with the dead in either of these two forms, especially if a ‘digital afterlife industry’ emerges to commercialize these. (shrink)

Sự xuật hiện của các công cụ AI tạo sinh trên nền tảng các mô hình ngôn ngữ lớn (LLMs) đã đem đến một công cụ mới cho con người, đặc biệt là trong các ngành sư phạm, báo chí, nhưng chúng cũng đem đến nhiều vấn đề Trong bài viết này, nhóm tác giả sẽ chỉ ra những những bất cập mới xuất hiện hoặc những vấn đề đã tồn tại nhưng có nguy cơ được đẩy lên cao hơn (...) do sự áp dụng của AI tạo sinh vào trong thực tế, tập trung vào hai nhóm vấn đề: công bằng xã hội và phổ cập tri thức khoa học và công tác truyền thông chính sách. Cụ thể, chúng tôi cho rằng các công cụ AI tạo sinh gặp khó khăn trong việc tiếp cận đến những nguồn kiến thức thuộc các văn hóa thiểu số, hoặc các ngôn ngữ nhỏ, có thể gây khó khăn trong việc tuyên truyền kiến thức khoa học, thực hiện chính sách. Tất cả những khó khăn này sẽ tạo ra những vấn đề tại Việt Nam và cần được giải quyết bằng việc phát triển một bộ giá trị, đạo đức và quy tắc ứng xử khi tương tác và sử dụng AI. (shrink)

Artificial agents create significant moral opportunities and challenges. Over the last two decades, discourse has largely focused on the concept of a ‘responsibility gap.’ We argue that this concept is incoherent, misguided, and diverts attention from the core issue of ‘control gaps.’ Control gaps arise when there is a discrepancy between the causal control an agent exercises and the moral control it should possess or emulate. Such gaps present moral risks, often leading to harm or ethical violations. We propose a (...) second-order ‘duty of moral control’ that mandates closing these gaps to reduce risks within acceptable moral limits. Our analysis encompasses both autonomous machines and collective agents, acknowledging their similarities and key differences in constitution and moral status. We suggest four methods to close control gaps: ensuring artificial agents attain moral agency, providing meaningful human control, implementing safety engineering, and employing social control. These methods aim to responsibly integrate artificial agents into society. We conclude that a realistic approach, which addresses the practical problems posed by control gaps, is essential. This approach provides solutions to manage the risks posed by artificial agents while maintaining acceptable moral standards, ensuring we responsibly harness their potential and address the ethical challenges they present. (shrink)

An AI-based “patient preference predictor” (PPP) is a proposed method for guiding healthcare decisions for patients who lack decision-making capacity. The proposal is to use correlations between sociodemographic data and known healthcare preferences to construct a model that predicts the unknown preferences of a particular patient. In this paper, I highlight a distinction that has been largely overlooked so far in debates about the PPP–that between algorithmic prediction and decision-making–and argue that much of the recent philosophical disagreement stems from this (...) oversight. I show how three prominent objections to the PPP only challenge its use as the sole determinant of a choice, and actually support its use as a source of evidence about patient preferences to inform human decision-making. The upshot is that we should adopt the evidential conception of the PPP and shift our evaluation of this technology towards the ethics of algorithmic prediction, rather than decision-making. (shrink)

This paper offers the first careful analysis of the possibility that AI and humanity will go to war. The paper focuses on the case of artificial general intelligence, AI with broadly human capabilities. The paper uses a bargaining model of war to apply standard causes of war to the special case of AI/human conflict. The paper argues that information failures and commitment problems are especially likely in AI/human conflict. Information failures would be driven by the difficulty of measuring AI capabilities, (...) by the uninterpretability of AI systems, and by differences in how AIs and humans analyze information. Commitment problems would make it difficult for AIs and humans to strike credible bargains. Commitment problems could arise from power shifts, rapid and discontinuous increases in AI capabilities. Commitment problems could also arise from missing focal points, where AIs and humans fail to effectively coordinate on policies to limit war. In the face of this heightened chance of war, the paper proposes several interventions. War can be made less likely by improving the measurement of AI capabilities, capping improvements in AI capabilities, designing AI systems to be similar to humans, and by allowing AI systems to participate in democratic political institutions. (shrink)

It is commonly agreed that so-called echo chambers and epistemic bubbles, associated with social media, are detrimental to liberal democracies. Drawing on John Rawls’s political liberalism, we offer a novel explanation of why social media platforms amplifying echo chambers and epistemic bubbles are likely contributing to the violation of the democratic norms connected to the ideal of public reason. These norms are clarified with reference to the method of (full) reflective equilibrium, which we argue should be cultivated as a civic (...) virtue in the context of political justification. The paper demonstrates how epistemic bubbles and echo chambers are likely to be detrimental to the disposition to follow the method of reflective equilibrium in the political context. Based on this diagnosis the paper highlights the need for interventions that help cultivate reflective equilibrium as a civic virtue and the need for corresponding interdisciplinary research. (shrink)

Privacy matters because it shields us from possible abuses of power. Human beings need privacy just as much as they need community. Our need for socialization brings with it risks and burdens which in turn give rise to the need for spaces and time away from others. To impose surveillance upon someone is an act of domination. The foundations of democracy quiver under surveillance. -/- This book is intended to contribute to a better understanding of privacy from a philosophical point (...) of view-what it is, what is at stake in its loss, and how it relates to other rights and values. The five parts that compose this book respond to five basic questions about privacy: Where does privacy come from? What is privacy? Why does privacy matter? What should we do about privacy? Where are we now? (shrink)

In recent years there has been a transition within society towards governments much more permeated by technology. One of the areas where this technological increase has been seen is in security processes. Currently, under the justification of a safer world, perpetual surveillance is given to the population through facial recognition cameras and data collection. For this reason, this research work seeks to explain how securitisation processes have become more complex to give way to much broader subjective prevention of security problems. (...) This research was carried out using a qualitative approach with a bibliographic review. Given these facts, it was found that this complex form of securitisation has led to transcending discursive acts and ending up institutionalising this securitisation. (shrink)

This booklet on Active Assisted Living (AAL) technologies has been created as part of the GoodBrother COST Action, which has run from 2020 to 2024. COST Actions are European research programs that promote collaboration across borders, uniting researchers, professionals, and institutions to address key societal challenges. GoodBrother focused on ethical and privacy concerns surrounding video and audio monitoring in care settings. The aim was to ensure that while AAL technologies help older adults and vulnerable individuals, their privacy and data protection (...) rights remain a top priority. This booklet is designed to guide you through the role that AAL technologies play in improving the quality of life for older adults, caregivers, and people with disabilities. AAL technologies offer tools for those facing cognitive or physical challenges. They can enhance independence, assist with daily routines, and promote a safer living environment. However, the rise of these technologies also brings important questions about data protection and user autonomy. (shrink)

Review of Joseph LeDoux: Das Netz der Persönlichkeit. Wie unser Selbst entsteht. Walter Verlag, Düsseldorf 2003. 510 Seiten (mit Abbildungen), 39,90 Euro. - Der eine Mensch ist mißtrauisch, der nächste leichtgläubig, diese ist warmherzig, jene kaltschnäuzig. Viele haben Charakter, manche sogar Persönlichkeit. Wie kommt es dazu? In seinem neuen Buch untersucht der Neurowissenschaftler Joseph LeDoux wie unser Selbst entsteht. In dem sehr lesbaren und angenehm übersetzten Werk wird anschaulich und detailliert berichtet, wie sich in unserem Gehirn die Charakteristika eines Individuums (...) ausbilden. (shrink)

This publication forms a basis for the interdisciplinary development of a certification system for artificial intelligence. In view of the rapid development of artificial intelligence with disruptive and lasting consequences for the economy, society, and everyday life, it highlights the resulting challenges that can be tackled only through interdisciplinary dialogue between IT, law, philosophy, and ethics. As a result of this interdisciplinary exchange, it also defines six AI-specific audit areas for trustworthy use of artificial intelligence. They comprise fairness, transparency, autonomy (...) and control, data protection as well as security and reliability while addressing ethical and legal requirements. The latter are further substantiated with the aim of operationalizability. (shrink)

Die vorliegende Publikation dient als Grundlage für die interdisziplinäre Entwicklung einer Zertifizierung von Künstlicher Intelligenz. Angesichts der rasanten Entwicklung von Künstlicher Intelligenz mit disruptiven und nachhaltigen Folgen für Wirtschaft, Gesellschaft und Alltagsleben verdeutlicht sie, dass sich die hieraus ergebenden Herausforderungen nur im interdisziplinären Dialog von Informatik, Rechtswissenschaften, Philosophie und Ethik bewältigen lassen. Als Ergebnis dieses interdisziplinären Austauschs definiert sie zudem sechs KI-spezifische Handlungsfelder für den vertrauensvollen Einsatz von Künstlicher Intelligenz: Sie umfassen Fairness, Transparenz, Autonomie und Kontrolle, Datenschutz sowie Sicherheit und (...) Verlässlichkeit und adressieren dabei ethische und rechtliche Anforderungen. Letztere werden mit dem Ziel der Operationalisierbarkeit weiter konkretisiert. (shrink)

We reply to Eva's comment on our "New Possibilities for Fair Algorithms," comparing and contrasting our Spanning criterion with his suggested Spacing criterion.

This paper explores the development of an ethical guardrail framework for AI systems, emphasizing the importance of customizable guardrails that align with diverse user values and underlying ethics. We address the challenges of AI ethics by proposing a structure that integrates rules, policies, and AI assistants to ensure responsible AI behavior, while comparing the proposed framework to the existing state-of-the-art guardrails. By focusing on practical mechanisms for implementing ethical standards, we aim to enhance transparency, user autonomy, and continuous improvement in (...) AI systems. Our approach accommodates ethical pluralism, offering a flexible and adaptable solution for the evolving landscape of AI governance. The paper concludes with strategies for resolving conflicts between ethical directives, underscoring the present and future need for robust, nuanced and context-aware AI systems. (shrink)

Generative AI technologies have the potential to be socially and politically transformative. In this paper, we focus on exploring the potential impacts that Generative AI could have on the functioning of our democracies and the nature of citizenship. We do so by drawing on accounts of deliberative democracy and the deliberative virtues associated with it, as well as the reciprocal impacts that social media and Generative AI will have on each other and the broader information landscape. Drawing on this background (...) theory, we outline some of the key positive and negative impacts that Generative AI is likely to have on democratic citizenship. The political significance of these impacts suggests the need for further regulation. (shrink)

This paper reviews the different meanings of vulnerability in the AI Act (AIA). We show that the AIA follows a rather established tradition of looking at vulnerability as a trait or a state of certain individuals and groups. It also includes a promising account of vulnerability as a relation but does not clarify if and how AI changes this relation. We spot the missing piece of the AIA: the lack of recognition that vulnerability is an inherent feature of all human-AI (...) interactions, varying in degree based on design choices and modes of interaction. Finally, we show how such a meaning of vulnerability may be incorporated into the AIA by interpreting the concept of “specific social situation” in Article 5 (b). (shrink)

This article presents a fairness principle for evaluating decision-making based on predictions: a decision rule is unfair when the individuals directly impacted by the decisions who are equal with respect to the features that justify inequalities in outcomes do not have the same statistical prospects of being benefited or harmed by them, irrespective of their socially salient morally arbitrary traits. The principle can be used to evaluate prediction-based decision-making from the point of view of a wide range of antecedently specified (...) substantive views about justice in outcome distributions. (shrink)

The now well-known impossibility results of algorithmic fairness demonstrate that an error-prone predictive model cannot simultaneously satisfy two plausible conditions for group fairness apart from exceptional circumstances where groups exhibit equal base rates. The results sparked, and continue to shape, lively debates surrounding algorithmic fairness conditions and the very possibility of building fair predictive models. This article, first, highlights three underlying points of disagreement in these debates, which have led to diverging assessments of the feasibility of fairness in prediction-based decision-making. (...) Second, the article explores whether and in what sense fairness as defined by the conjunction of the implicated fairness conditions is (un)attainable. Drawing on philosophical literature on the concept of feasibility and the role of feasibility in normative theory, I outline a cautiously optimistic argument for the _diachronic feasibility_ of fairness. In line with recent works on the topic, I argue that fairness can be _made possible_ through collective efforts to eliminate inequalities that feed into local decision-making procedures. (shrink)

Algorithmic risk assessment tools, such as COMPAS, are increasingly used in criminal justice systems to predict the risk of defendants to reoffend in the future. This paper argues that these tools may not only predict recidivism, but may themselves causally induce recidivism through self-fulfilling predictions. We argue that such “performative” effects can yield severe harms both to individuals and to society at large, which raise epistemic-ethical responsibilities on the part of developers and users of risk assessment tools. To meet these (...) responsibilities, we present a novel desideratum on algorithmic tools, called explainability-in-context, which requires clarifying how these tools causally interact with the social, technological, and institutional environments they are embedded in. Risk assessment practices are thus subject to high epistemic standards, which haven't been sufficiently appreciated to date. Explainability-in-context, we contend, is a crucial goal to pursue in addressing the ethical challenges surrounding risk assessment tools. (shrink)

I consider statistical criteria of algorithmic fairness from the perspective of the _ideals_ of fairness to which these criteria are committed. I distinguish and describe three theoretical roles such ideals might play. The usefulness of this program is illustrated by taking Base Rate Tracking and its ratio variant as a case study. I identify and compare the ideals of these two criteria, then consider them in each of the aforementioned three roles for ideals. This ideals program may present a way (...) forward in the normative evaluation of candidate statistical criteria of algorithmic fairness. (shrink)

The increasing use of artificial intelligence (AI) systems in our daily life through various applications, services, and products explains the significance of trust/distrust in AI from a user perspective. AI-driven systems have significantly diffused into various fields of our lives, serving as beneficial tools used by human agents. These systems are also evolving to act as co-assistants or semi-agents in specific domains, potentially influencing human thought, decision-making, and agency. Trust/distrust in AI plays the role of a regulator and could significantly (...) control the level of this diffusion, as trust can increase, and distrust may reduce the rate of adoption of AI. Recently, varieties of studies have paid attention to the variant dimension of trust/distrust in AI, and its relevant considerations. In this systematic literature review, after conceptualization of trust in the current AI literature review, we will investigate trust in different types of human-Machine interaction, and its impact on technology acceptance in different domains. In addition to that, we propose a taxonomy of technical (i.e., safety, accuracy, robustness) and non-technical axiological (i.e., ethical, legal, and mixed) trustworthiness metrics, and some trustworthy measurements. Moreover, we examine some major trust-breakers in AI (e.g., autonomy and dignity threat), and trust makers; and propose some future directions and probable solutions for the transition to a trustworthy AI. (shrink)

We introduce a fairness criterion that we call Spanning. Spanning i) is implied by Calibration, ii) retains interesting properties of Calibration that some other ways of relaxing that criterion do not, and iii) unlike Calibration and other prominent ways of weakening it, is consistent with Equalized Odds outside of trivial cases.

In ‘Rawlsian algorithmic fairness and a missing aggregation property of the difference Principle’, the authors argue that there is a false assumption in algorithmic fairness interventions inspired by John Rawls’ theory of justice. They argue that applying the difference principle at the level of a local algorithmic decision-making context (what they term a ‘constituent situation’), is neither necessary nor sufficient for the difference principle to be upheld at the aggregate level of society at large. I find these arguments compelling. They (...) are in line with long-standing concerns about Rawlsian justice in general, and its application in ‘fair’ machine learning in particular. Overall, this article is a very welcome contribution to the literature. In this commentary, I would like to briefly highlight some limitations of these arguments, and some of their broader implications for the field of political philosophy-inspired fair machine learning. (shrink)

It is commonly accepted that clinicians are ethically obligated to disclose their use of medical machine learning systems to patients, and that failure to do so would amount to a moral fault for which clinicians ought to be held accountable. Call this ‘the disclosure thesis.’ Four main arguments have been, or could be, given to support the disclosure thesis in the ethics literature: the risk-based argument, the rights-based argument, the materiality argument and the autonomy argument. In this article, I argue (...) that each of these four arguments are unconvincing, and therefore, that the disclosure thesis ought to be rejected. I suggest that mandating disclosure may also even risk harming patients by providing stakeholders with a way to avoid accountability for harm that results from improper applications or uses of these systems. (shrink)

Dwork, Hardt, Pitassi, Reingold, & Zemel [6] introduced two notions of fairness, each of which is meant to formalize the notion of similar treatment for similarly qualified individuals. The first of these notions, which we call additive metric fairness, has received much attention in subsequent work studying the fairness of a system composed of classifiers which are fair when considered in isolation [3, 4, 7, 8, 12] and in work studying the relationship between fair treatment of individuals and fair treatment (...) of groups [6, 7, 13]. Here, we extend these lines of research to the second, less-studied notion, which we call multiplicative metric fairness. In particular, we exactly characterize the fairness of conjunctions and disjunctions of multiplicative metric fair classifiers, and the extent to which a classifier which satisfies multiplicative metric fairness also treats groups fairly. This characterization reveals that whereas additive metric fairness becomes easier to satisfy when probabilities of acceptance are small, leading to unfairness under functional and group compositions, multiplicative metric fairness is better-behaved, due to its scale-invariance. (shrink)

There are calls from various sectors, including the popular press, industry, and academia, to incorporate artificial intelligence (AI)-based technologies in general, and large language models (LLMs) (such as ChatGPT and Gemini) in particular, into various spheres of the South African higher education sector. Nonetheless, the implementation of such technologies is not without ethical risks, notably those related to bias, unfairness, privacy violations, misinformation, lack of transparency, and threats to autonomy. This paper gives an overview of the more pertinent ethical concerns (...) that may result from the deployment of various current AI technologies in the South African higher education context. It provides a broad overview of the relevant AI ethics literature and distills a few general AI-ethics principles that can serve as guidelines for the ethical development, adoption, and use of AI systems. Suggestions are made as to how these might be applied to mitigate the ethical concerns in the South African higher education context. Overall, it is argued that AI technologies should only be adopted if they offer demonstrable benefits to stakeholders affected by them and that care should be taken to ensure that any potential harms are adequately addressed. (shrink)

This paper considers the ethical risks and opportunities presented by generalist medical artificial intelligence (GMAI), a kind of dynamic, multimodal AI proposed by Moor et al. (2023) for use in health care. The research objective is to apply widely accepted principles of biomedical ethics to analyze the possible consequences of GMAI, while emphasizing the distinctions between GMAI and current-generation, task-specific medical AI. The principles of autonomy and health equity in particular provide useful guidance for the ethical risks and opportunities of (...) novel AI systems in health care. The ethics of two applications of GMAI are examined: enabling decision aids that inform and educate patients about certain treatments and conditions, and expanding AI-driven diagnosis and treatment recommendation. Emphasis is placed on the potential of GMAI to improve shared decision-making between patients and providers, which supports patient autonomy. Another focus is on health equity, or the reduction of health and access disparities facing underserved populations. Although GMAI presents opportunities to improve patient autonomy, health literacy, and health equity, premature or inadequately regulated adoption of GMAI has the potential to compromise both health equity and patient autonomy. On the other hand, there are significant risks to health equity and autonomy that may arise from not adopting GMAI that has been thoroughly validated and tested. A careful balancing of these risks and benefits will be required to secure the best ethical outcome, if GMAI is ever employed at scale. (shrink)

The proliferation of Artificial Intelligence (AI) has sparked an overwhelming number of AI ethics guidelines, boards and codes of conduct. These outputs primarily analyse competing theories, principles and values for AI development and deployment. However, as a series of recent problematic incidents about AI ethics/ethicists demonstrate, this orientation is insufficient. Before proceeding to evaluate other professions, AI ethicists should critically evaluate their own; yet, such an evaluation should be more explicitly and systematically undertaken in the literature. I argue that these (...) insufficiencies could be mitigated by developing a research agenda for a feminist metaethics of AI. Contrary to traditional metaethics, which reflects on the nature of morality and moral judgements in a non-normative way, feminist metaethics expands its scope to ask not only what ethics is but also what our engagement with it should be like. Applying this perspective to the context of AI, I suggest that a feminist metaethics of AI would examine: (i) the continuity between theory and action in AI ethics; (ii) the real-life effects of AI ethics; (iii) the role and profile of those involved in AI ethics; and (iv) the effects of AI on power relations through methods that pay attention to context, emotions and narrative. (shrink)