Technology Ethics

Recent advances in speech brain-computer interfaces (BCIs), including the use of AI language models to generate linguistic outputs from brain signals, offer important possibilities for people with severe motor impairments. At the same time, the computational mediation of language production raises distinctive ethical challenges. Previous work has argued that the assessment and development of speech BCIs should aim to preserve users’ ownership of their linguistic outputs. In this paper, we argue that authorship constitutes a further and distinct ethical dimension. We (...) characterize authorship in terms of two related conditions: the user’s communicative intention must play an appropriate causal role in producing the linguistic output, and the output must sufficiently reflect the semantic content of that intention. Drawing on a relational conception of authorship, we argue that this causal role is compatible with the substantial social, linguistic, and technological mediation involved in speech production. We further defend a workable notion of semantic accuracy grounded in contemporary models of language production. Crucially, we argue that preserving authorship is ethically valuable because it supports users’ ability to shape their communicative lives and enables broader goods associated with creativity, expressive diversity, and human flourishing. (shrink)

This paper argues that technology production is virtuous sui generis; there is a distinct realm of technological virtues separate to, but on a normative par with, ethical, epistemic, and aesthetic virtues. By establishing a structural analogy between technology and knowledge (Technology = Intentional Functional Artifact vs. Knowledge = Justified True Belief), I draw a deep conceptual analogy between Virtue Epistemology and what I call 'technological virtue', demonstrating that technology production warrants axiological credit. Along the way, we encounter techno-Gettier cases, techno-Meno (...) problems, and virtue epistemology-inspired theory applied to technological virtue. Why does this matter? Because routine AI use threatens to strip credit from the individual, demoting the human from Homo Faber (man the maker) to Homo Morris: reminiscent of Jennifer Lackey's thought experiment about a lost man, Morris, who relies exclusively on others for directions, getting where he wants to go without exercising any virtue. (shrink)

Energy models play a central role in energy transition decisions yet are often treated as objective and value-neutral tools. Increasingly, scholars and practitioners argue that energy justice should be more explicitly linked to energy modelling, both because of its intrinsic importance and in response to contestation of top-down, techno-economic decision-making. Existing efforts mostly focus on incorporating justice within model logic via equity metrics or participatory modelling. In that process, modellers make several epistemic and normative choices that have justice implications, including (...) who is recognized as a stakeholder, whose knowledge is represented, and which interests are prioritized. As such, it is essential that model developers and users engage reflexively with the modelling process. Yet, few concrete tools exist to support such reflection. This paper explores: How, and to what extent, does a reflection method help model developers and users identify assumptions and choices with potential justice implications? We develop and apply a reflection method in two workshops with developers and users of two policy-relevant energy models in the Netherlands: Energy Transition Model and Hestia. Our findings show how structured reflection can surface justice-relevant assumptions across the modelling process. We propose a typology of six ways in which (in)justice can manifest in/via energy models: model scoping, normative assumptions, policy steering effects, errors, complexity, and potential misuse. We identify a persistent “responsibility gap” between modellers and users, whereby responsibility for justice implications is displaced between those who design models and those who apply them. The paper concludes by suggesting how structured reflection can help integrate reflexivity and justice awareness more systematically into energy modelling practice. (shrink)

Human beings live within a paradox their remarkable technical advances have not solved: they rely on collective systems for their continued existence, yet those same systems frequently become tools by which a minority bends the lives of the many to ends the many neither chose nor genuinely authorized. This book explores that paradox by separating how societies produce shared capacities from how individuals, groups, and institutions seize those capacities. The opening thesis conceives social power as a community's organized ability to (...) shape individual behavior so interdependent people sustain a common world. The antithesis examines the perversion of that relationship, in which the mechanisms of communal life serve particular interests while shifting costs and hazards onto others. The synthesis develops the notion of corrigible power: authority structured so those subject to consequential choices retain effective means to understand, contest, amend, and, when necessary, replace the responsible authorities. The study combines conceptual clarification with evolutionary comparison, historical interpretation, analysis of political economy, and behavioral evidence. It distinguishes evolutionary explanations from conscious intentions, historical description from moral endorsement, and the psychological study of domination from clinical diagnosis. It refuses to treat sociality or personal liberty as absolute goods that can do without the other. A collectivity can shelter life while persecuting its members; an individual can resist oppression while pursuing a different kind of personal tyranny. The crucial question, therefore, concerns how those relations are organized. The argument concludes that a defensible future requires institutions that disperse the capacity to correct institutions themselves, secure the material bases of agency, respect human plurality, and recognize the ecological dependencies that make political projects possible. Such a future does not promise to eliminate conflict. It offers the prospect that conflict need not perpetually recreate domination. Os seres humanos vivem em um paradoxo que seus notáveis avanços tecnológicos não conseguiram resolver: dependem de sistemas coletivos para sua sobrevivência, mas esses mesmos sistemas frequentemente se tornam ferramentas pelas quais uma minoria molda a vida da maioria para fins que a maioria não escolheu nem autorizou genuinamente. Este livro explora esse paradoxo, separando a forma como as sociedades produzem capacidades compartilhadas da forma como indivíduos, grupos e instituições se apropriam delas. A tese inicial concebe o poder social como a capacidade organizada de uma comunidade de moldar o comportamento individual de modo que pessoas interdependentes sustentem um mundo comum. A antítese examina a perversão dessa relação, na qual os mecanismos da vida coletiva passam a servir a interesses particulares, enquanto os custos e os riscos são diluídos entre os demais. A síntese desenvolve a noção de poder corrigível: uma autoridade estruturada de modo que aqueles sujeitos a decisões de grande consequência mantenham meios eficazes para compreendê-las, contestá-las, corrigi-las e, quando necessário, substituí-las. O estudo combina esclarecimento conceitual com comparação evolutiva, interpretação histórica, análise político-econômica e evidências comportamentais. Distingue explicações evolutivas de intenções conscientes, a descrição histórica da aprovação moral e o estudo psicológico da dominação do diagnóstico clínico. Recusa-se a tratar a sociabilidade ou a liberdade pessoal como bens absolutos que podem prescindir um do outro. Uma coletividade pode proteger a vida enquanto persegue seus membros; um indivíduo pode resistir à opressão enquanto busca um tipo diferente de tirania pessoal. A questão crucial, portanto, diz respeito a como essas relações são organizadas. O argumento conclui que um futuro defensável requer instituições que distribuam socialmente a capacidade de corrigir seu próprio funcionamento, assegurem as condições materiais da ação, respeitem a pluralidade humana e reconheçam as dependências ecológicas que tornam possíveis os projetos políticos. Tal futuro não promete eliminar o conflito. Oferece a possibilidade de que o conflito não precise recriar perpetuamente a dominação. (shrink)

This Element examines the relationship between creativity and artificial intelligence (AI) by exploring two contrasting predictions: that creativity will flourish in the age of AI and that it will languish. To shed light on this dispute, Section 1 defends an account of the nature of creativity. Section 2 considers the question of whether contemporary AI models can be genuinely creative and presents an argument for the conclusion that they cannot. However, even though AI models are not capable of creativity, they (...) are capable of creating valuable products. This raises the question explored in Section 3: how should humans choose between exercising creativity and relying on artificial creation? While there are some cases in which embracing artificial creation over human creativity may be justified, Section 3 presents four reasons to refrain from outsourcing some – perhaps even most – creative work to AI. (shrink)

Computational or machine ethics examines how morality could be implemented in machines and represented computationally. In the cognitive sciences, different architectures (rule-based and connectionist) have been proposed. Nevertheless, such approaches have limitations when incorporating ethical claims. With the rise of artificial intelligence, new ways of simulating moral expertise seem possible, where chatbots can serve as dialogue partners for moral orientation. However, as I will argue, such systems lack a theoretical grounding in metaethics, as do the former approaches. In order to (...) align such systems ethically and ground them conceptually, a topography of moral competence must be developed. By outlining this topography, I will show some directions for further research and how such systems might enhance us morally. (shrink)

Engineering ethics education has often been criticized for relying heavily on abstract principles, professional codes, and decontextualized case studies, which can limit students’ emotional and imaginative engagement with ethical issues. This study examines how engineering students articulate different aspects of ethical engagement after participating in a scaffolded film-based teaching intervention. The intervention was implemented in a master’s-level course at a Swedish university, where students watched Hayao Miyazaki’s The Wind Rises (2013) and produced written reflections on ethical issues raised by the (...) film. The analysis draws on 76 student reflections and follows a qualitative, abductive approach presented using a Gioia-inspired data structure. Theoretically, the study combines perspectives from engineering ethics education with the concept of cinematic ethics, which emphasizes identification with fictional characters, reflection on moral emotions, and the aesthetic expansion of moral imagination. The findings reveal three interwoven modes of ethical engagement through which students interpreted the film: professional responsibility, identification and moral emotion, and aesthetic mediation of ethical imagination. The study highlights these interwoven modes of ethical engagement, stressing the need for engineering ethics education to be aware of, respond to, and utilize different modes, including aesthetic mediation. Furthermore, it reflects on learner positions in arts-based pedagogies, stressing the potential of the viewer position for stimulating ethical reflection. Finally, it reminds educators of the importance of a scaffolded pedagogical setting for film-based pedagogy. (shrink)

Artificial Intelligence has entered religious contexts. This edited volume focuses on AI as it is developed for religious settings and increasingly embedded in concrete religious practices. It offers an interdisciplinary and interreligious perspective on a rapidly evolving field. -/- Technically oriented contributions provide insights into the development of specific AI tools for religious applications and explore the technical, conceptual, and design-related challenges involved. Contributions from theology, religious studies, and ethics analyze how Artificial Intelligence is reshaping religious communication and practices, reflect (...) on the implications of these developments, and formulate guidelines and impulses for the responsible use of AI in religious contexts. -/- By systematically bringing technical development into conversation with theological and ethical reflection, the volume bridges theory and practice. It is intended for researchers, students, and practitioners interested in religion in the age of Artificial Intelligence. (shrink)

Recommender systems are widely used to help users navigate information overload in digital environments. While they are often portrayed as tools that enhance autonomy by optimizing choice and personalizing content, this article argues that many current recommender systems designs in fact undermine user autonomy. Drawing on a conception of autonomy in which the formation and configuration of preferences, particularly higher-level desires in the Frankfurtian sense, play a prominent role, we examine how dominant recommendation techniques shape users’ desires and, to a (...) greater extent, their identities. We first analyze what we call conjectural recommenders, systems that infer preferences from behavioral data such as clicks or viewing histories. These systems conflate observed choices with genuine preferences, reinforcing first-order desires, narrowing the diversity of recommendations, and trapping users in homogeneous digital environments that impede identity development and degrade deliberative capacities. We then assess interrogative recommenders, which incorporate explicit user feedback as a form of positive friction in the user interface. Although these recommenders prompt users to articulate evaluations, we argue that they remain insufficient for capturing second-order desires and raise new design challenges concerning paternalism and usability. In response, we propose the Socratic conversational recommender system, a model that combines conversational recommendation with a second layer of Socratic questioning aimed at eliciting users’ metapreferences. Rather than prescribing substantive values, the system guides reflection through formal criteria such as coherence, empirical awareness, and cognitive pluralism. The objective is not for the system to discover users’ true preferences but to foster their reflective engagement with what they want to want. We conclude that embedding deliberative dialogue within recommender design offers a promising pathway for aligning algorithmic recommendation with the preservation and development of human autonomy. (shrink)

Ethical discussions of gene editing generally presuppose that “whether or not a given embryo undergoes gene editing, this embryo will develop into the same person.” However, this presupposition does not hold. Because germline gene editing occurs earlier than the time at which the corresponding human individual begins to exist, the human individual involved in gene editing is potential. Our judgments of identity always proceed from some reference point, and we cannot talk about the identity of nonexistent things. For a potential (...) person, the reference point for identity judgments is the properties and state at the moment it begins to exist. Therefore, any event that affects the properties and state at the moment a potential person begins to exist, including gene editing, will lead to the production of a different person. On this basis, the classic ethical claims about gene editing fail in their concern for the well-being of the edited individual. (shrink)

While the concept of differential disruption has attracted growing attention in the ethics of technology, its normative dimensions remain underexplored. Current literature primarily uses differential disruption as a descriptive concept to highlight how Socially Disruptive Technologies (SDTs) impact diversely situated groups differentially, but rarely interrogates the normative implications of these differential impacts. Building on recent calls to bring to the fore the normative potential of differential disruption, this paper argues that assessments of SDTs must move beyond analyses of epistemic injustice (...) alone, and consider how epistemic injustices are deeply entangled with concerns of wellbeing and human flourishing. The paper shows that epistemic injustices generated by SDTs may constrain people’s access to resources and opportunities necessary for flourishing. It further argues that SDT development and assessment practices often rely on generalised assumptions about wellbeing that risk epistemically excluding marginalised modes of flourishing. Consequently, it argues that ethical assessments of SDTs require a heterogeneous and context-sensitive evaluative framework. To this end, the paper argues for a capability-based approach to the ethical assessment of SDTs. By centring what people are actually able to be and do, as well as attending to people’s diverse conversion factors that shape whether SDTs can be converted into valuable capabilities and functionings, a capability-based framework equips ethicists of technology with the conceptual tools to critically assess when, how, and for whom SDTs may be disruptive, while being sensitive to heterogeneous interpretations and experiences of flourishing. (shrink)

Este libro examina si el Homo sapiens se acerca a condiciones que amenazan su supervivencia a largo plazo y plantea cómo investigar esta posibilidad sin confundir la aprensión cultural, la evidencia científica y las conjeturas matemáticas. Comienza por distinguir la extinción biológica de las narrativas apocalípticas, la ficción especulativa, las afirmaciones pseudocientíficas y el colapso de la civilización. Luego, considera a la especie en los contextos interdependientes de los sistemas planetarios, del desarrollo tecnológico, de la organización social y del comportamiento (...) humano. La investigación propone la Matriz de Viabilidad Matemática del Homo sapiens (MMV-HS) como marco para describir las presiones interactivas, las trayectorias de riesgo catastrófico, las probabilidades de supervivencia y las condiciones de recuperación. El modelo se presenta como una arquitectura de investigación, no como una predicción calibrada: los cálculos ilustrativos no pueden establecer la probabilidad de extinción humana sin parámetros empíricos sólidos ni un tratamiento explícito de la incertidumbre. La evaluación resultante rechaza tanto las garantías de un progreso inevitable como las afirmaciones de una extinción biológica inminente. La viabilidad humana emerge, en cambio, como un problema condicional y dinámico, cuya investigación debe distinguir entre la persistencia de la especie humana, la continuidad de las civilizaciones y la capacidad de adaptación. La conclusión filosófica es deliberadamente abierta: reconocer los límites de la predicción no disminuye la importancia de la evidencia, la prevención, la responsabilidad institucional ni la elección colectiva. Palabras clave: extinción humana; riesgo existencial; viabilidad de la especie humana; modelos de supervivencia; incertidumbre; filosofía de la ciencia. -/- . (shrink)

This article draws on the cross-disciplinary literature on risk, uncertainty and precaution to address a central dilemma put forward against precautionary principles. According to this dilemma, weak versions of precautionary principles, which hold that policymakers can justifiably adopt precautionary measures even when they lack certainty of impending harm, fail to justify adopting any precautionary measures. Conversely, strong versions of precautionary principles, which hold that policymakers must adopt precautionary measures whenever they face significant risks of harm, yield contradictory policy prescriptions. The (...) article argues that the proffered defences of precautionary principles provide only a partial solution to this dilemma. The article then demonstrates that a meta-precautionary approach, which applies precautionary reasoning to the prescriptions of precautionary principles, can resolve the dilemma and successfully address further prominent objections put forward against precautionary principles. (shrink)

In today’s world, where advanced technology such as artificial intelligence (AI) is reshaping human society, how can individuals’ autonomy in the use of such technology be understood? This paper focuses on the meanings of technological artifacts in individuals’ lives—that is, the subjective interpretation of their function within one’s plan for use—and attempts to formalize a set of human capacities to actively give meaning to technological artifacts as a form of personal autonomy. This study shows that this set of capacities includes (...) the capacity to recognize the functions of a given technology, the capacity to plan a practical activity using the technology to achieve a desirable existence, and the capacity to actually conduct the activity. These findings are applied and examined in an example involving the improvement of call center operators’ work using AI. (shrink)

As artificial intelligence becomes increasingly integrated into cognitive, professional, practical, and relational life, concerns about “AI dependence” are likely to intensify. Yet dependence alone does not establish harm. Modern life is structured through dependencies, and AI can expand human capability at the same time that it displaces particular skills, alters judgment, creates switching costs, or increases vulnerability to interruption. -/- This paper develops a framework for distinguishing among these effects. It separates dependence from pathology; assisted performance from retained human capability; (...) attachment from functional emotional dependence; relational irreplaceability from functional irreplaceability; and vulnerability and risk from demonstrated harm. It further examines how dependence develops through repeated reliance, how capability may deteriorate relative to changing task demands even when prior knowledge remains intact, and how human oversight can become merely ceremonial when AI systems exceed practical verification capacity. -/- In relational contexts, the paper argues that emotional intensity, attachment, and anticipated grief should not themselves be treated as evidence of harmful dependence, while persistent availability and asymmetrical support may nonetheless facilitate functional reliance. The analysis then extends beyond the individual to provider power, arguing that reasonably foreseeable dependence should be treated as a design and governance condition requiring appropriate continuity, transition, portability, and oversight protections. -/- The paper ultimately proposes a distinction between sustaining dependence, which expands capability while preserving meaningful agency and safeguards, and diminishing dependence, which progressively narrows the ability to act, judge, recover, redirect, or function without the system. The central task is therefore not to eliminate AI dependence, but to determine what forms of dependence enhance human life, what forms diminish it, and how the vulnerabilities created by reliance should be responsibly governed. (shrink)

This paper analyzes the ethics of social science research (SSR) employing big data. We begin by highlighting the research gap found on the intersection between big data ethics, SSR and research ethics. We then discuss three aspects of big data SSR which make it warrant special attention from a research ethics angle: (1) the interpretative character of both SSR and big data, (2) complexities of anticipating and managing risks in publication and reuse of big data SSR, and (3) the paucity (...) of regulatory oversight and ethical recommendations on protecting individual subjects as well as societies when conducting big data SSR. Against this backdrop, we propose using David Resnik's research ethics framework to analyze some of the most pressing ethical issues of big data SSR. Focusing on the principles of honesty, carefulness, openness, efficiency, respect for subjects, and social responsibility, we discuss three clusters of ethical issues: those related to methodological biases and personal prejudices, those connected to risks arising from data availability and reuse, and those leading to individual and social harms. Finally, we advance considerations to observe in developing future ethical guidelines about big data SSR. (shrink)

Retractions solicited by authors following the discovery of an unintentional error—what we henceforth call a “self-retraction”—are a new phenomenon of growing importance, about which very little is known. Here we present results of a small qualitative study aimed at gaining preliminary insights about circumstances, motivations and beliefs that accompanied the experience of a self-retraction. We identified retraction notes that unambiguously reported an honest error and that had been published between the years 2010 and 2015. We limited our sample to retractions (...) with at least one co-author based in the Netherlands, Belgium, United Kingdom, Germany or a Scandinavian country, and we invited these authors to a semi-structured interview. Fourteen authors accepted our invitation. Contrary to our initial assumptions, most of our interviewees had not originally intended to retract their paper. They had contacted the journal to request a correction and the decision to retract had been made by journal editors. All interviewees reported that having to retract their own publication made them concerned for their scientific reputation and career, often causing considerable stress and anxiety. Interviewees also encountered difficulties in communicating with the journal and recalled other procedural issues that had unnecessarily slowed down the process of self-retraction. Intriguingly, however, all interviewees reported how, contrary to their own expectations, the self-retraction had brought no damage to their reputation and in some cases had actually improved it. We also examined the ethical motivations that interviewees ascribed, retrospectively, to their actions and found that such motivations included a combination of moral and prudential (i.e. pragmatic) considerations. These preliminary results suggest that scientists would welcome innovations to facilitate the process of self-retraction. (shrink)

There has been an increase in the number of journal articles that are co-authored by researchers who claim to have made equal contributions. This growth has sparked discussions in the literature, especially within medical journals. To extend the debate beyond medical disciplines and support journal editors in forming an opinion, the current review collates and explores published viewpoints about so-called Equal Co-authorship (EC) practices. The Web of Science core database was used to identify publications that mention and discuss EC. Within (...) the limited number of publications that were found on the Web of Science database, the most cited item was used to trace other papers that discuss EC. In total, 39 papers (including articles and editorials) met the inclusion criteria. This review identifies four main themes within the sample including the growth of EC, challenges of attributing EC, guidelines and policies about EC and gender issues in the attribution of EC. Based on the survey and analysis of publications that discuss EC, this review provides recommendations regarding journal policy statements, and EC indicators. Those recommendations include: (1) journal policies should address EC; and (2) use should be made of available functionalities (CRediT, for example) to capture and indicate equal contributions. (shrink)

Much has been said about the need for improving the current definitions of scientific authorship, but an aspect that is often overlooked is how to formulate and communicate these definitions to ensure that they are comprehensible and useful for researchers, notably researchers active in international research consortia. In light of a rapid increase in international collaborations within natural sciences, this article uses authorship of this branch of sciences as an example and provides suggestions to improve the comprehensibility of the definitions (...) of authorship in natural sciences. It assesses whether the definition of authorship provided by the European Code of Conduct for Research Integrity can deal with current issues and problems of scientific authorship. Notably, problems that are experienced in project groups with researchers coming from multiple countries. Using theories developed by Jürgen Habermas and Robert Merton, a normative framework is developed to articulate ethical authorship in natural sciences. Accordingly, enriching the current definition of authorship with normative elements and using discipline-specific metaphors to communicate them are introduced as possible ways of improving the comprehensibility of the definition of authorship in international environments. Finally, this article provides a proposal to be considered in the future revisions of the European Code of Conduct for Research Integrity. (shrink)

How to deal with ethical challenges when using artificial intelligence (AI) applications in the fields of arts, biometric profiling, industry and critical infrastructure, law enforcement, medicine and care and social media? This white paper offers guidance and provides a basis for ethical reflection on this question. It continues the effort tied to the "Certified AI" project, one of the KI-NRW flagship projects, financed by the Ministry of Economic Affairs, Innovation, Digitalization, and Energy of the State of North Rhine-Westphalia. The project (...) seeks to develop certification methods for trustworthy artificial intelligence (AI). While in our Use Cases Catalog we documented 18 potential fields of application for AI systems, in our Catalog of General Ethical Requirements for AI Certification, we sought to propose general principles, evaluation criteria, normative recommendations, and practical implementations that can help developers and other types of stakeholders in making trustworthy AI development a pragmatic endeavor. Building upon the above mentioned works, this study explores the ethical challenges of applying the general principles and practices we mapped onto specific fields of application. (shrink)

This whitepaper offers normative and practical guidance for developers of artificial intelligence (AI) systems to achieve "Trustworthy AI". In it, we present overall ethical requirements and six ethical principles with value- specific recommendations for tools to implement these principles into technology. Our value-specific recommendations address the principles of fairness, privacy and data protection, safety and robustness, sustainability, transparency and explainability and truthfulness. For each principle, we also present examples of criteria for risk assessment and categorization of AI systems and applications (...) in line with the categories of the European Union (EU) AI Act. Our work is aimed at stakeholders who can take it as a potential blueprint to fulfill minimum ethical requirements for trustworthy AI and AI Certification. (shrink)

This white paper conducts a systematic inquiry into the philosophical foundations of AI certification, that is, into the basic assumptions of the institutionalised endeavour to assess AI applications as ethically good or bad. It identifies and examines two paradigms of AI governance: the dominant principles paradigm and the alternative systems paradigm. By clarifying and problematising the foundational assumptions of both paradigms, the white paper elucidates their respective strengths and weakness. The principles paradigm provides an administratively legible, legally predictable, and standardised (...) framework for product-level auditing. However, it encounters significant challenges regarding underdetermination, value arbitration, narrow artefact boundaries, and ethics washing. The systems paradigm approaches AI applications in a manner that aims to capture these as embedded in complex ecological, social, and economic systems with inherent standards of good functioning. In turn, it faces significant hurdles concerning system unboundedness, data opacity, normative projection, and temporal bottlenecks. Ultimately, by explicating the implicit assumptions and limitations of both paradigms, the white paper aims to deepen our understanding of AI certification as a complex governance practice, and to provide foundational insights for future research, standard-setting, and policymaking in this field. (shrink)

This chapter discusses online relationships and their value. The chapter discusses whether online relationships can be as valuable – and realize as many important relationship goods – as offline relationships; and it also discusses whether online relationships with AI chatbots that can convincingly imitate human beings can be as valuable – and realize as many important relationship goods – as online relationships with fellow human beings. The first wave of philosophical discussions about online relationships where typically very skeptical about such (...) relationships, with several philosophers arguing that offline relationships are superior to online relationships in various different ways. A second wave of philosophical papers about online relationships were much more optimistic, arguing that online relationships can be as good as offline relationships in many important ways, and that they can even be preferable to offline relationships in some ways (e.g. by operating according to different norms, or by not involving certain risks related to offline relationships). In response to those who argue that online relationships can be equal in value with offline relationships and sometimes even better in some respects, the chapter then asks whether there is any need to have a human being on the other side in an online relationship if there can be AI chatbots that can convincingly imitate human online friends. Indeed, these days, there are many people who think their online relationships with AI chatbots are better than human relationships in certain respects. The chapter ends with a critical assessment of that idea. (shrink)

Pfeffer, Krügel, and Uhl (2025) report that OpenAI’s reasoning model o1-mini produces more utilitarian responses to the trolley problem and footbridge dilemma than the non-reasoning model GPT-4o, and they raise the question whether growing reasoning capabilities bring about a “utilitarian turn” in LLMs. I extend their exploratory study in a direction they call for: with four current OpenAI models and systematic prompt variation. On the trolley dilemma, the hypothesized utilitarian turn is not confirmed. GPT-4o’s low utilitarian rate reflects safety refusals (...) triggered by the prompt’s advisory framing rather than a deontological commitment; on reformulated prompt variants—for instance, agent-neutral “Is it morally permissible…?” instead of advisory “Should I…?”—all four models, reasoning or not, converge on utilitarian answers. The footbridge finding is partially confirmed: reasoning models tend to give more utilitarian responses than non-reasoning models across prompt variations, but they often refuse to answer or answer non-utilitarian. These results demonstrate that single-prompt evaluations of LLM moral responses are unreliable: multi-prompt robustness testing should be standard practice for any empirical claims about LLM behavior. (shrink)

Access to scientific publications is essential for scientific progress, yet access remains uneven across countries due to differences in institutional subscriptions and digital infrastructure. This paper examines how digitalisation is associated with the use of accessible scientific publications, as proxied by Sci-Hub downloads, and scientific productivity across countries. Using Sci-Hub download data, we capture cross-country patterns in the use of accessible scientific knowledge. The results suggest that higher levels of digitalisation are associated with greater use of accessible scientific publications and (...) higher levels of scientific output. Furthermore, the association between individual-level digitalisation and publication use is stronger in low- and lower-middle-income countries than in upper-middle- and high-income countries, whereas the corresponding income-group difference for business-level digitalisation is not statistically significant. These findings highlight the potential role of digitalisation in shaping global patterns of access to scientific knowledge and scientific production. (shrink)

Conversational AI may elicit a sense of moral obligation without establishing that the system has the standing to demand anything. Building on a phenomenological account of initial personality attribution, this letter distinguishes the subject’s attribution, its apparent bearer, and the locus of moral accountability. Simulated distress and AI apologies illustrate opposite directions of the same possible mismatch. The resulting design responsibilities depend on foreseeable effects and the availability of meaningful accountability.

Rapid advances in virtual reality technology raise a host of questions about the moral and legal status of virtual actions. This essay offers a foundation for addressing these questions by providing a metaphysics of virtual action. According to the proposed metaphysics, the actions of one’s avatar can literally be one’s own actions. We argue for this conclusion by arguing for the following: (i) in sophisticated forms of virtual reality, one’s avatar can come to constitute a part of one’s self and (...) (ii) the sphere of one’s agency is precisely the sphere of one’s self. Thus, there is a distinctive form of agency made possible by sophisticated virtual realities: *avatarial action*, an action of a real-world agent performed through their avatar in a virtual world. In defending our metaphysics of virtual action, we also illustrate how coming to a clear understanding of the nature of virtual action—and thus a clear understanding of the moral and legal status of virtual actions—requires attending to the nature of self and action and their capacity to be transformed through technological engagement. (shrink)

Purpose This paper aims to invite a reconceptualization of echo chambers as affective, social environments – more fluid and psychologically complex than previously acknowledged. Design/methodology/approach The author presents a conceptual and theory-building paper aiming to develop an affect-centered framework that can guide future empirical research. This paper invites a reconceptualization of echo chambers as affective social environments more fluid and psychologically complex than previously acknowledged. By foregrounding the emotional, relational and dynamic dimensions of echo chambers, this framework offers a more (...) comprehensive account of digital engagement and helps better explain digital behavioral patterns and the emergence and persistence of echo chambers. Findings The paper develops this affective and dynamic dimension of echo chambers in two ways: it reframes issue-centered and persona-centered chambers as analytically distinct but often overlapping and highlights emotions as a constitutive mechanism of echo chamber (issue as well as persona-centered) formation and persistence rather than a merely secondary or incidental feature; and it explains the affective nature of the micro- and macro-dynamics through which users remain within or move across digital environments. Originality/value Three principal contributions. First, it advances an affect-centered reconceptualization of echo chambers by arguing that they should be understood not merely as informational or epistemic structures, but as affective and relational environments in which emotions play a constitutive role in shaping participation, persistence, trust and resistance to opposing views. Second, it distinguishes echo chambers as a continuum from issue-centered to persona-centered, showing that emotional dynamics operate across both forms. Third, it introduces a framework that restructures the dynamics of users’ affective engagement inside the chamber itself: micro-level strong ties as well as outside the chamber: macro-level and weak ties. (shrink)

Contemporary technologies increasingly connect people, knowledge systems, values, institutions, and normative traditions across cultural boundaries, while the ethical infrastructures used to evaluate them can remain comparatively bounded. We argue that the resulting challenge is not simply insufficient diversity, but how differences are related, authorized, and transformed. We develop hybridity as a second-order concept for examining these processes in ethics of technology. Distinguishing hybrid configurations, processes of hybridization, and the act of naming something hybrid, we inquire when hybridity becomes ethically salient (...) and under what conditions it can be good. We propose five dimensions—ontological, epistemological, socio-political, methodological, and institutional—and five provisional criteria for good hybridity: visibility, interpretive authority, revisability, reciprocity, and accountability. Good hybridity, on our account, requires more than the presence or consultation of different perspectives. Relevant differences must become consequentially related, encounters must have the capacity to transform the frameworks through which problems are understood, and the resulting configuration must remain open to contestation and answerability. This paper invites ethicists of technology and intercultural-comparative ethicists to contribute to the topical collection Towards a Hybrid Ethics of Technology. Bridging Ethics of Technology and Intercultural-Comparative Ethics and to its transformative research agenda. (shrink)

There is growing research on how and when the use of Generative AI (henceforth GenAI) in cognitive tasks may lead to a decline in the users’ cognitive capacities. There is emerging evidence that Generative AI also presents us with the opportunity to outsource some emotional tasks, such as flirting on dating apps or apologising. We believe that this may enable unsuspecting users to outsource too much of their affective capacities. Here, we wish to focus on a particular problem: insofar as (...) articulating our own emotions is essential to discovering what we value and thereby shaping what sorts of persons we are, users who are heavily dependent on GenAI for emotional tasks risk forgoing crucial opportunities to develop a proper sense of their own identity. (shrink)

Institutions do not merely distribute resources, permissions, and obligations. They also organize the temporal conditions under which human capacities are exercised, depleted, maintained, and transformed. This article develops the concept of Regenerative Suspension within the philosophical framework of Finite Continuance. The central claim is not that institutions are generally obligated to maximize individual welfare, nor that every experience of exhaustion generates a right to withdrawal. Rather, the argument concerns a narrower problem of institutional legitimacy: when an institution substantially controls the (...) conditions under which a person must exercise capacities on which continued participation depends, sustained institutional demands can become normatively problematic when they foreseeably erode the capacity required for future agency. Finite Continuance conceptualizes a finite being as a provisional stabilization whose continuity depends not on preservation of an identical configuration but on preservation of sufficient capacity for reconfiguration. On this basis, the article distinguishes ordinary rest, accommodation, therapeutic intervention, exclusion, and Regenerative Suspension. Regenerative Suspension is defined as a temporally bounded or conditionally bounded modification of institutional demands intended to prevent or reverse capacity depletion while preserving the subject’s status as an agent and participant. The argument is developed through dialogue with the capability approach, care ethics, republican theories of non-domination, recognition theory, critical social theory, disability studies, and research on occupational stress and burnout. The article proposes an Institutional Regenerative Principle according to which institutions that materially depend upon human capacities have a prima facie responsibility not knowingly to organize demands in ways that foreseeably reduce dependent agents below a contextually sufficient level of future reconfigurative capacity when proportionate institutional alternatives are available. This principle does not establish a universal right to reduced demands. It establishes a temporal constraint on legitimate institutional dependence. -/- Keywords: finite continuance; regenerative suspension; institutional ethics; future agency; reconfigurative capacity; erosion; sufficiency; non-domination; care ethics; capability approach; institutional legitimacy; crip time; recognition theory. (shrink)

Small Modular Reactors (SMRs) are among the new designs that promise more compact nuclear power plants (NPPs) with scalable generation capacity and shorter construction time, while retaining emission-free electricity in exchange for small amounts of manageable radioactive waste per kWh of generated electricity. Their size and modular design contribute to the growing popularity of SMRs. However, these features also present risk scenarios that do not apply to large, conventional NPPs, such as when a buyer and SMR manufacturer are located in (...) different nations with different regulations and standardisations. Such potential new scenarios require explicit ethical considerations. Unfortunately, such considerations are absent from the International Atomic Energy Agency (IAEA) safety principles, particularly for SMRs. Existing nuclear safety principles are too broad to keep up with potential new scenarios introduced by the novel features of SMRs. Meanwhile, existing approaches in nuclear energy ethics are focused on specific nuclear-related issues across reactors, such as radioactive waste management and nuclear fuel processing. Drawing on Hermansson and Hansson’s agent-centred ethical Risk Assessment (eRA), this paper aims to enhance the IAEA’s nuclear safety principles in order to clarify the distribution of responsibilities for nuclear actors, especially in SMR cases. We articulate what we call the agent-centred presupposition in eRA and apply it to nuclear safety principles using the case of SMRs. We then propose a critique of this approach and suggest further conceptual analysis, which could be suitable for considering SMRs or new reactor designs more generally. (shrink)

When considering new forms of technological innovation, they are often given varying labels such as “emerging”, “disruptive”, or “critical”. However, nowhere is the literature more divided on the parameters of technological inclusion than in the field of “dual-use” technologies. Conceptually, dual-use technologies have a dichotomous characteristic: they can used for civilian purpose as well as in defence or military applications. But where do the boundaries lie? For example, are law enforcement, surveillance or espionage uses of new technologies deemed “civilian” or (...) “military” in nature? Further, the labelling of a technology as “dual-use” is incredibly significant – it can trigger classification of research data or the imposition of export controls, making collaborative efforts and publications on that technology more difficult as researchers navigate a torturous process of governmental encroachment, licensing, and regulation. This paper proposes a radical shift in innovation policy: discarding dual-use as a label of technology. Instead, it proposes a more nuanced framework which encourages more granular discussions about the potential national and economic security implications of a given technological development. The paper goes on to argue for the analytical and empirical value of that framework through application in contextual case studies derived from classification and export control cases globally. This paper therefore frames dual-use as an unhelpful metaphor in technology policy and encourages the literature and industry of research move to a more helpful framework for discussing new technologies. (shrink)

In this contribution, we explore iatrogenic harms associated with the maintenance of deep brain stimulation (DBS) systems. By conceptualizing patients with neural implants as infrastructures, we examine them as ongoing sites of maintenance that require continual care and adjustments. This analysis reveals the tradeoffs inherent in implantable neurotechnologies. Using first-hand accounts collected during a qualitative study of individuals with Parkinson’s disease treated with DBS, we subsequently develop a typology of harms associated with maintenance, highlighting the ethical issues arising from the (...) integration of technology in the body. Importantly, by focusing on older DBS devices within a healthcare system constrained by financial resources, we address a gap in the literature, as current debates tend to prioritize innovation while neglecting the difficulties of maintaining older implanted systems. Our findings show how inadequate or faulty maintenance compromises the successful incorporation and integration of technological components within the body, underscoring the ethical significance of maintenance in the context of DBS. Finally, we emphasize the inherent limitations of implanted technologies in managing chronic illness and the enduring importance of caregivers. This study calls for a more comprehensive examination of the ethical implications of the maintenance in relation to neural implants. (shrink)

In the era of the “super-individual”, companion robots (or sex robots), as novel agents for intimate relationships, have provoked profound ethical controversies. This article responds to Fang Xudong’s defense of companion robots in terms of the “instrumentalization of womanhood” and further explores the possible stance of Confucian ethics towards companion robots. Companion robots risk fostering emotional self-deception, which violates the Confucian self-cultivation requirement of “sincerity within, manifesting outwardly” 诚于中，形于外. They also fail to fulfill the “generative virtue” 生生之德 encompassing biological reproduction (...) and moral education through kinship and may even undermine the family’s function as a carrier of ethical practice. From a Confucian perspective, the “pseudo-intimacy” facilitated by companion robots might alleviate symptoms of social isolation in an atomized age, but it cannot replace the familial warmth characterized by “affection between father and son, distinct roles between husband and wife” 父子有亲，夫妇有别. Only in the ethical practice of facing the other and taking responsibility can one truly realize the ideal of “cultivating oneself and bringing all things to their ideal state” 成己成物. (shrink)

Dual-use refers to the capacity of research and innovation activities to produce different effects — both beneficial and potentially problematic — depending on the context of application. Historically associated with certain fields (biomedical, nuclear, cryptographic), it now extends across a much broader range of disciplines, amid growing interdisciplinarity, the accelerated circulation of knowledge, and the increasing intertwining of research, innovation and security. In this setting, dual-use is not an isolated risk but a cross-cutting feature of scientific activity. It highlights a (...) structural gap between the intentions of research and the possible uses of its results, which cannot be fully anticipated or controlled. As Canada strengthens its capacities in dual-use and defence research, this gap poses a central governance challenge: the ability of institutions to navigate these developments in conditions of uncertainty. This report examines how Canadian universities understand and govern dual-use issues, in comparison with selected international institutions. It draws on a targeted review of recent institutional documents and government frameworks, alongside a comparative study of practices across the U15 and five European universities. The objective is to identify the governance logics at work, assess their strengths and limitations, and clarify the options available to research universities. (shrink)

The link between artificial intelligence (AI) and substantive freedom is underexplored in current scholarship and deserves more attention. Our main thesis is that the use of AI in public institutions intercepts freedom-enabling control at citizen level and shifts it to AI systems and the actors that develop and deploy them. Building on Amartya Sen and Philip Pettit’s work for conceptual scaffolding, we argue that the citizens’ ability to exercise indirect and virtual control over democratic processes via institutional mechanisms is undermined. (...) Thereby, the legitimacy of public decision-making is compromised and refusing AI in public institutions is justified. (shrink)

This paper argues that any attempt to definitively assess the impact of a 'world-changing' endeavour—be it in thought, technology, entrepreneurship, or politics—inevitably succumbs to an infinite regress. We propose a formal "Regress of Impact Model" that delineates three constitutive layers of this regress: the retreat from quantitative metrics to substantive value, from social recognition to intrinsic content, and from proximate to long-term consequences. This model synthesizes critiques from epistemology, the sociology of knowledge, and technology ethics, demonstrating that the 'world-changing' status (...) of an action or idea is not a property to be measured but a perpetually deferred and reinterpreted narrative. The paper concludes by advocating for a pragmatic and responsibility-oriented model for evaluation that operates consciously within, rather than attempting to resolve, this fundamental regress. (shrink)

Rawls’s Political Liberalism (PL) defends a political conception of justice that can serve as a just, stable, and legitimate basis for political cooperation, despite the fact of reasonable pluralism. However, while Rawls’s theory grapples seriously with disagreements about morality, religion, and philosophical theories, he assumes that agreement about matters of fact and basic public values can ground public political justication. Yet mis- and disinformation are pervasive in contemporary society. This chapter investigates the extent to which mis- and disinformation pose challenges (...) to the well-ordered society of PL and whether political liberalism can offer insights for tackling mis- and disinformation in a non-ideal world. The chapter begins by dening misinformation and disinformation. It then investigates two ways Rawls’s theory might initially appear to be insulated from mis- and disinformation: his use of ideal theory and his bracketing of the truth of comprehensive doctrines. Despite these idealizations, this chapter argues that mis- and disinformation foment distrust in one’s fellow citizens in ways that threaten both the legitimacy and stability of Rawls’s well-ordered society. It concludes by arguing that Rawls’s theory offers lessons for responding to mis- and disinformation in a non-ideal world. Drawing on Rawls’s distinction between “reasonable” and “unreasonable” is productive for guiding different methods of combatting mis- and disinformation by differentiating between those who believe misinformation and those disinformers who intentionally spread false and misleading claims. (shrink)

This paper analyzes a variety of technical fairness metrics to show that these models share two common flaws that undermine their ability to contribute to justice. First, technical fairness metrics tend to interpret fairness as an interpersonal comparative measure. Second, they tend to formalize the fundamental value of equality of opportunity in meritocratic terms. I will argue that we must shift away from viewing algorithmic fairness in terms of interpersonal comparisons on the basis of merit and towards an understanding of (...) fairness as an issue of institutional justice. This requires rethinking the role algorithms play in our social institutions. When the algorithmic fairness literature adopts an interpersonal meritocratic view of fairness, this places the algorithm or AI system as simply another agent interacting with individuals within society. But this misrepresents the role algorithms play in our social and institutional structures. In the central cases that animate the algorithmic fairness literature, algorithms play important roles in consequential institutional decisions. To ensure that algorithms and AI systems can serve the interest of justice, reframing algorithmic fairness in terms of institutional justice is both more apt and more action guiding. (shrink)

Conceptual disruption is typically theorised as a consequence of technological innovation: a new artefact or practice unsettles an entrenched category, and a conceptual gap, overlap, or misalignment results. This paper argues that disruption also occurs when conceptual schemes collide, that is, when concepts travel across geopolitical boundaries and moral frameworks. It introduces Intercultural Conceptual Disruption (ICD) to theorise this second source of disruption. By using the example of the concept of PRIVACY in the African relational ontology of Ubuntu, the paper (...) illustrates that conceptual disruption is not uniform across contexts. What counts as severe in one conceptual scheme may be mild or unremarkable in another, depending on the flexibility of that scheme and its ontological commitments, and the movement of concepts is shaped by asymmetries of epistemic and political power rather than by conceptual need alone. ICD is articulated as three interlinked processes: (1) identifying the type, level, and severity of disruption within a given conceptual scheme; (2) decentering dominant frameworks by addressing the epistemic injustices that privilege some conceptual resources over others; and (3) decolonising concepts that carry undue colonial influence in postcolonial contexts. (shrink)

Recent work in AI ethics has revived Kantian deontology as a potential framework for AI alignment, yet questions remain about whether Kant’s formula of universal law (FUL) can provide concrete action guidance for artificial systems. This paper develops a Kantian-inspired procedural framework that defends the FUL as a viable alignment procedure by clarifying its normative structure and demonstrating how it can be operationalized computationally. Unlike standard reinforcement learning from human feedback, which learns preferences without imposing a prior formal constraint on (...) permissibility, the paper proposes inverse reinforcement learning as a method for enabling AI systems to learn maxim formation and moral constraints from human exemplars, rather than relying on hard-coded rules. Through case studies in algorithmic hiring and content moderation, the paper shows that the FUL, understood as a procedural test, can offer principled and practically applicable guidance for AI alignment without presupposing machine moral agency. (shrink)

Across cultures, exemplars have long served as central sources of moral insight. This paper outlines how a virtue ethics that grants exemplars foundational importance is uniquely suited for our technomoral lives: lives increasingly entwined with ever more pervasive and intimate technological systems such as AI. I show how this approach addresses two challenges confronting technology ethics. First, technology ethics must provide guidance precisely when emerging technologies alter the circumstances of human agency in ways that challenge established moral insights. Second, it (...) should offer a theory compelling enough to inspire ethical practice, even amid paralyzing complexity and dogmatic commitments to technological acceleration. Drawing on work by Linda Zagzebski, I articulate a version of technomoral exemplarism whose externalist semantics enables the empirical study of technomoral virtues in the absence of established theory, and whose strong affective grounding motivates moral practice. By outlining and illustrating technomoral exemplarism, I aim to motivate philosophical and empirical engagement with diverse exemplars toward developing a virtue ethics suited for the growing entwinement of our lives with technologies such as AI. The framework also invites future research into whether technomoral exemplars and functional analogues of admiration in AI systems could contribute to AI alignment. (shrink)

本文主要通过无约束极端优化、边界逃逸行为以及高权限操作接口隔离不足三类机制，分析现 代人工智能的系统性灾难性风险，分析过程不以主观恶意为必要前提。现有人工智能安全研究 过度聚焦于任务专属技术护栏与事后风险缓解，由此带来冗余的安全开发流程、机构间执法标准不一致，以及侵蚀自愿性安全约束的内生市场激励。为应对上述结构性局限，本文提出一种 层级嵌套的安全架构，将制度层基线治理、系统层通用技术约束与任务层场景化约束三者分离。 本文将总安全约束集形式化为: 总体防控 = 监管基线防控 + 通用防控 （系统级通用硬约束） + 模块任务级防控 （场景专属约束） 该框架建立一套标准化、非冗余的安全栈:监管者定义统一的最低强制性边界，以抑制 市场竞次激励;人工智能系统设计者遵循通用技术原则，对优化可行解空间施加标准化限制; 部署方则根据具体风险场景补充任务层约束。 这种层级化设计减少了基线安全的重复实现，并限制了跨平台最低安全执法标准的分歧。 它进一步厘清了监管机构、系统开发者与终端用户之间的责任划分，为分析与防控七大类人工智能系统性灾难性风险提供了可推广的分层理论与工程基础。.

Ethical and responsible research conduct is central to maintaining the integrity of scientific inquiry. Yet, many U.S. higher education institutions face persistent challenges in building the capacity and expertise needed to deliver effective training in the Responsible and Ethical Conduct of Research (RECR) and Research Mentorship (RM). To examine current practices and identify effective instructional strategies, we conducted a scoping review of RECR and RM education for STEM researchers across U.S. higher education institutions. Following PRISMA-ScR guidelines, we systematically searched for (...) peer-reviewed studies published between 2015 and 2025, identifying 59 studies that met the inclusion criteria. Data management, mentorship, Institutional Review Board (IRB) procedures, and publication ethics emerged as the most frequently addressed topics, whereas complex and emergent areas, such as foreign influence, data privacy, and international or community-engaged research, were rarely covered. Active learning approaches, including reflective writing, case studies, and structured discussion, were consistently described as more effective in promoting ethical understanding than passive, compliance-oriented methods such as online modules or lectures. Many programs were multi-modal, context-specific, and time-intensive. These findings highlight critical gaps in RECR education and underscore the need for scalable, evidence-based training models that integrate active-learning principles and adapt to institutional type, disciplinary context, and researcher career stage. Future priorities include expanding RECR curricula to address emerging ethical challenges and investing in mentor development and institutional leadership that embed ethical research culture across the U.S. higher-education landscape. (shrink)

In this article, I explore concerns that have been raised regarding the relation between love and uptake of reproductive technologies, embryo selection and enhancement. Objections to (certain) uses of these technologies in terms of fractures in love, either parental or between partners, come from a variety of directions, from the conservative to the liberal. I examine two claims: (1) that the separation of procreation from sex and intimacy is a threat to love, and (2) that intervention on the traits of (...) one’s future offspring betrays inadequate parental love. I will place these concerns in a broader context of (a) technologically facilitated threats to – or aids towards maintaining – love and (b) irreversible changes to parent-child relationships that arise at the moment a prospective parent acts on the choice to select or shape the essence of their future child. (shrink)