Ethics of Artificial Intelligence

Recently, Simon Goldstein and Cameron Domenico Kirk-Giannini have argued that certain existing AI systems are capable of well-being. They consider the three leading approaches to well-being—hedonism, desire satisfactionism, and the objective list approach—and argue that theories of these kinds plausibly imply that some current AI systems are capable of welfare. In this paper, I argue that the leading versions of each of these theories do not imply this. I conclude that we have strong reason to doubt that current AI systems (...) are capable of well-being. (shrink)

This paper investigates how human interactions with AI-powered chatbots may offend human dignity. Current chatbots, driven by large language models (LLMs), mimic human linguistic behaviour but lack the moral and rational capacities essential for genuine interpersonal respect. Human beings are prone to anthropomorphise chatbots—indeed, chatbots appear to be deliberately designed to elicit that response. As a result, human beings’ behaviour toward chatbots often resembles behaviours typical of interaction between moral agents. Drawing on a second-personal, relational account of dignity, we argue (...) that interacting with chatbots in this way is incompatible with the dignity of users. We show that, since second-personal respect is premised on reciprocal recognition of second-personal moral authority, behaving towards chatbots in ways that convey second-personal respect is bound to misfire in morally problematic ways, given the lack of reciprocity. Consequently, such chatbot interactions amount to subtle but significant violations of self-respect—the respect we are dutybound to show for our own dignity. We illustrate this by discussing four actual chatbot use cases (information retrieval, customer service, advising, and companionship), and propound that the increasing societal pressure to engage in such interactions with chatbots poses a hitherto underappreciated threat to human dignity. (shrink)

Governments and social scientists are increasingly developing machine learning methods to automate the process of identifying terrorists in real time and predict future attacks. However, current operationalizations of “terrorist”’ in artificial intelligence are difficult to justify given three issues that remain neglected: insufficient construct legitimacy, insufficient criterion validity, and insufficient construct validity. I conclude that machine learning methods should be at most used for the identification of singular individuals deemed terrorists and not for identifying possible terrorists from some more general (...) class, nor to predict terrorist attacks more broadly, given intolerably high risks that result from such approaches. (shrink)

Peter Königs recently argued that, while autonomous robots generate responsibility gaps, such gaps need not be considered problematic. I argue that Königs’ compromise dissolves under analysis since, on a proper understanding of what “responsibility” is and what “gap” (metaphorically) means, their joint endorsement must repel an attitude of indifference. So, just as “calamities that happen but don’t bother anyone” makes no sense, the idea of “responsibility gaps that exist but leave citizens and ethicists unmoved” makes no sense.

This paper categorizes current approaches to AI ethics into four perspectives and briefly summarizes them: (1) Case studies and technical trend surveys, (2) AI governance, (3) Technologies for AI alignment, (4) Philosophy. In the second half, we focus on the fourth perspective, the philosophical approach, within the context of applied ethics. In particular, the explicability of AI may be an area in which scientists, engineers, and AI developers are expected to engage more actively relative to other ethical issues in AI.

Beauty filters are automated photo editing tools that use artificial intelligence and computer vision to detect facial features and modify them, allegedly improving a face’s physical appearance and attractiveness. Widespread use of these filters has raised concern due to their potentially damaging psychological effects. In this paper, I offer an account that examines the effect that interacting with such filters has on self-perception. I argue that when looking at digitally-beautified versions of themselves, individuals are looking at AI-curated distorted mirrors. This (...) helps identify two potential cognitive effects of this behavior. First, it can elicit affective attitudes that change how individuals feel when looking at their unfiltered self-images. Second, exposure to filtered self-images might cause a perceptual normalization of such images. Finally, I argue that this form of distorted mirror gazing is a novel cultural practice for self-perception, and I highlight some ways in which this practice could be critically evaluated and ultimately changed. (shrink)

Griefbots are chatbots modeled on the personalities of deceased individuals, designed to assist with the grieving process and, according to some, to continue relationships with loved ones after their physical passing. The essay examines the promises and perils of griefbots from a Daoist perspective. According to the Daoist philosopher Zhuangzi, death is a natural and inevitable phenomenon, a manifestation of the constant changes and transformations in the world. This approach emphasizes adaptability, flexibility, and openness to alternative ways of relating to (...) the deceased after they are gone. Drawing on the Daoist ideas, this essay distinguishes between two importantly different ways to continue a relationship with the deceased, and correspondingly, two kinds of griefbot designs: preservational griefbots and transformational griefbots. A preservational griefbot, which aims to preserve the user’s relationship with the deceased, presents itself as the deceased individual by replicating the deceased’s personality as closely and for as long as possible. By contrast, a transformational griefbot aligns more with Zhuangist ideals and aims to facilitate the transformation of our relationship with the deceased. It helps the bereaved come to terms with their loss and move on with their lives, without abandoning their relationship with the deceased. Specifically, it facilitates this transformation in two ways: first, by helping the bereaved release themselves from their habitual ways of relating to the loved one, thereby bringing closure to their relationship in its previous form; and second, by enabling the bereaved to continue interacting with the ideas and legacies of the deceased in a new and digitally mediated way, which, in turn, reshapes and enriches the content and significance of their relationship. (shrink)

The emergence of artificially sentient beings raises moral, political, and legal issues that deserve scrutiny. First, it may be difficult to understand the well-being elements of artificially sentient beings and theories of well-being may have to be reconsidered. For instance, as a theory of well-being, hedonism may need to expand the meaning of happiness and suffering or it may run the risk of being irrelevant. Second, we may have to compare the claims of artificially sentient beings with the claims of (...) humans. This calls for interspecies aggregation, which is a neglected form of interpersonal aggregation. Lastly, there are practical problems to address, such as whether to include artificially sentient beings in the political decision-making processes, whether to grant them a right to self-determination in digital worlds, and how to protect them from discrimination. Given these, the emergence of artificially sentient beings compels us to reevaluate the positions we typically hold. (shrink)

This paper aims to provide a roadmap for governing AI. In contrast to the reigning paradigms, we argue that AI governance should be not merely a reactive, punitive, status-quo-defending enterprise, but rather the expression of an expansive, proactive vision for technology—to advance human flourishing. Advancing human flourishing in turn requires democratic/political stability and economic empowerment. To accomplish this, we build on a new normative framework that will give humanity its best chance to reap the full benefits, while avoiding the dangers, (...) of AI. This new framework of “Power-Sharing Liberalism” is a philosophy that restores protections of positive liberties to liberalism. As we deploy it here, it helps shape a more comprehensive (and we would contend, more accurate) understanding of both risk and opportunity introduced by AI. To lay out how Power-Sharing Liberalism can be applied to AI governance, we take four steps. First, we define central concepts in the field of AI governance, disambiguating between forms of technological harms and risks. Second, we review current normative frameworks around the globe and argue that Power-Sharing Liberalism is a better fit for governing AI. Third, we walk through the six governance tasks that should be accomplished by any governance framework and analyze them through the Power-Sharing Liberalism framework. Based on that analysis, we make 17 recommendations for the governance of AI—including transformative investments in public goods, personnel, and the sustainability of democracy itself. Finally, we discuss concrete proposals for implementing those recommendations. (shrink)

This piece in Nature Medicine notes the risks that incorporation of AI systems into health care poses to disabled patients and proposes ways to avoid them and instead create benefit.

Este manual sobre tecnologías para un envejecimiento activo y saludable, también conocido como Vida Asistida Activa (Active Assisted Living – AAL en sus siglas en inglés), ha sido creado como parte de la Acción COST GoodBrother, que se ha llevado a cabo desde 2020 hasta 2024. Las Acciones COST son programas de investigación europeos que promueven la colaboración internacional, uniendo a investigadores, profesionales e instituciones para abordar desafíos sociales importantes. GoodBrother se ha centrado en las cuestiones éticas y de privacidad (...) relacionadas con la monitorización empleando cámaras y micrófonos en entornos de provisión de cuidados. El objetivo ha sido garantizar que, mientras las tecnologías AAL ayudan a las personas mayores y a los individuos vulnerables, sus derechos de privacidad y protección de datos se mantengan como una prioridad. Este manual está diseñado para guiar en el papel que juegan las tecnologías AAL en la mejora de la calidad de vida de personas mayores, sus cuidadores y personas con discapacidades. Las tecnologías AAL ofrecen herramientas para quienes enfrentan desafíos cognitivos o físicos. Pueden mejorar la independencia, asistir en las rutinas diarias y promover un entorno de vida más seguro. Sin embargo, el aumento de estas tecnologías también plantea preguntas importantes sobre la protección de datos y la autonomía del usuario. A~través de estudios de casos prácticos, este manual explora escenarios reales en los que se aplican las tecnologías AAL. Ofrecen información sobre sus beneficios y desafíos. Se discuten situaciones en las que personas mayores han recuperado su sentido de independencia, por ejemplo, mediante el uso de sistemas de detección de caídas o dispositivos de monitorización de salud. Estas historias ofrecen una perspectiva humana de la tecnología, mostrando cómo pueden mejorar el bienestar mientras plantean cuestiones sobre la privacidad y el uso ético. El manual también destaca la importancia de tomar decisiones informadas al elegir e implementar sistemas AAL. Se enfatiza la~necesidad de que estas tecnologías equilibren seguridad y conveniencia con respeto a los derechos individuales. Encontrará información sobre cuestiones clave, incluyendo cómo seleccionar dispositivos adecuados, gestionar la configuración de privacidad y garantizar que la tecnología se integre de manera fluida en la vida diaria. Además de estos casos de estudio, el manual aborda temas importantes, como la seguridad de los datos, las consideraciones éticas y el papel de las normativas europeas, como el Reglamento General de Protección de Datos (RGPD), en la protección de la privacidad del usuario. Comprender estos aspectos es esencial para asegurar que los sistemas AAL no solo sean efectivos, sino también respetuosos de los derechos y la dignidad de los usuarios. Este recurso está destinado a una amplia audiencia, incluidos usuarios finales, cuidadores, profesionales de la salud y responsables de políticas. Proporciona orientación práctica sobre cómo integrar las tecnologías AAL en entornos de cuidados mientras se protege la privacidad y se garantiza un uso ético. Las ideas presentadas aquí buscan empoderar a los usuarios y cuidadores para que tomen decisiones informadas que mejoren tanto la calidad del cuidado como el respeto por la autonomía personal. (shrink)

Many social media platforms enable (nearly) anyone to post (nearly) anything. One clear downside of this permissiveness is that many people appear bad at determining who to trust online. Hacks, quacks, climate change deniers, vaccine skeptics, and election deniers have all gained massive followings in these free markets of ideas, and many of their followers seem to genuinely trust them. At the same time, there are many cases in which people seem to reliably determine who to trust online. Consider, for (...) example, Do It Yourself (DIY) content about how to play guitar, bake, fix one’s plumbing, or repair one’s car. For these topics, those who have the largest accounts and the most popular content typically possess significant expertise. That is, social media users seem to reliably pick out DIY experts. We thus have a puzzle: why are social media users seemingly competent at identifying DIY experts, but not climate science or vaccine experts? In what follows, we solve this puzzle. We begin by identifying a novel wisdom of the crowds phenomenon: specifically we argue that the crowd (in combination with social media search and recommendation algorithms) reliably picks out DIY experts and serves as a credentialing institution for DIY content. Next, we argue that (a) there are five epistemic factors that determine whether the crowd can succeed at recognizing experts on social media platforms, and (b) while many of those factors are satisfied to a sufficiently high degree by DIY content, they are mostly lacking for content about climate change or vaccines. (shrink)

Imagine that you are given access to an AI chatbot that compellingly mimics the personality and speech of a deceased loved one. If you start having regular interactions with this “thanabot,” could this new relationship be a continuation of the relationship you had with your loved one? And could a relationship with a thanabot preserve or replicate the value of a close human relationship? To the first question, we argue that a relationship with a thanabot cannot be a true continuation (...) of your relationship with a deceased loved one, though it might support one’s continuing bonds with the dead. To the second question, we argue that, in and of themselves, relationships with thanabots cannot benefit us as much as rewarding and healthy intimate relationships with other humans, though we explain why it is difficult to make reliable comparative generalizations about the instrumental value of these relationships. (shrink)

Abstract:This article examines the incorporation of applied ethics into artificial intelligence (AI) within Chilean university curricula, emphasizing the urgent need to implement an integrated framework of action. Through a documentary analysis, it becomes evident that most higher education programs do not explicitly include AI ethics courses in their curricula, highlighting the need for institutionalizing this integration systematically. In response, we propose an approach grounded in feminist science and data feminism, advocating for the inclusion of diverse perspectives and experiences in the (...) teaching of applied ethics. This framework aims not only to enhance the integration of ethics into the curriculum but also to prepare students to address ethical dilemmas in complex sociotechnical contexts, underscoring the importance of applied ethical reasoning in AI-related disciplines. -/- Keywords: AI ethics, professional ethics, sociotechnics, feminist epistemology, diversity. -/- Resumen: Este artículo examina la incorporación de la ética aplicada a la inteligencia artificial (IA) en los currículos universitarios chilenos, destacando la urgencia de implementar un marco de acción integrado. Mediante un análisis documental, se evidencia que la mayoría de los programas de educación superior no declaran cursos de ética en IA en sus currículos, lo que alerta sobre la necesidad de sistematizar esta integración institucionalmente. En respuesta, proponemos un enfoque basado en el feminismo en la ciencia y el feminismo de datos, promoviendo la inclusión de diversas perspectivas y experiencias en la enseñanza de la ética aplicada. Este marco busca mejorar la integración de la ética en el currículo y también preparar a los estudiantes para resolver dilemas éticos en contextos sociotécnicos complejos, reforzando la necesidad del razonamiento ético aplicado en la formación en disciplinas asociadas a la IA. -/- Palabras clave: ética de IA, ética profesional, sociotecnología, epistemología feminista, diversidad. (shrink)

Virtual reality is not only a prowess of technological advancement and AI, but also an element that extends the horizons of human existence and complicates the way of approaching various phenomena of the physical world, for example, violence. Its practice in virtuality leads to a series of challenges, especially when virtual reality is considered as genuine reality. This text delves into virtual violence, the influence of AI on it and the problems that its conception implies. To analyze this phenomenon, parallax (...) is proposed as a concept that helps to understand its development. This theory invites us to see violence in different ways, thus influencing the vision of morality in the virtual world. Although these reflections outline difficulties, a position is proposed that invites to urgently address the need for an applicable morality to virtual and physical fields, emphasizing dignity as the purpose of human practices. (shrink)

The convergence of humans and artificial intelligence (AI) systems introduces new dynamics into the cultural and intellectual landscape. Complementing emerging cultural evolution concepts such as machine culture, AI agents represent a significant techno-sociological development, particularly within the anthropological study of Web3 as a community focused on decentralization through blockchain. Despite their growing presence, the cultural significance of AI agents remains largely unexplored in academic literature. Toward this end, we conceived hybrid netnography, a novel interdisciplinary approach that examines the cultural and (...) intellectual dynamics within digital ecosystems by analyzing the interactions and contributions of both human and AI agents as co-participants in shaping narratives, ideas, and cultural artifacts. We argue that, within the Web3 community on the social media platform X, these agents challenge traditional notions of participation and influence in public discourse, creating a “hybrid marketplace of ideas”—a conceptual space where human- and AI-generated ideas coexist and compete for attention. We examine the current state of AI agents in idea generation, propagation, and engagement, positioning their role as cultural agents through the lens of memetics and encouraging further inquiry into their cultural and societal impact. Additionally, we address the implications of this paradigm for privacy, intellectual property, and governance, highlighting the societal and legal challenges of integrating AI agents into the hybrid marketplace of ideas. (shrink)

This paper examines how professionals (N = 32) working on artificial intelligence (AI) view structural AI ethics challenges like injustices and inequalities beyond individual agents' direct intention and control. This paper answers the research question: What are professionals’ perceptions of the structural challenges of AI (in the agri-food sector)? This empirical paper shows that it is essential to broaden the scope of ethics of AI beyond micro- and meso-levels. While ethics guidelines and AI ethics often focus on the responsibility of (...) designers and the competencies and skills of designers to take this responsibility, our results show that many structural challenges are beyond their reach. This result means that while ethics guidelines and AI ethics frameworks are helpful, there is a risk that they overlook more complicated, nuanced, and intersected structural challenges. In addition, it highlights the need to include diverse stakeholders, such as quadruple helix (QH) participants, in discussions around AI ethics rather than solely focusing on the obligations of AI developers and companies. Overall, this paper demonstrates that addressing structural challenges in AI is challenging and requires an approach that considers four requirements: (1) multi-level, (2) multi-faceted, (3) interdisciplinary, and (4) polycentric governance. (shrink)

We introduce the Bias Network Approach (BNA) as a sociotechnical method for AI developers to identify, map, and relate biases across the AI development process. This approach addresses the limitations of what we call the "isolationist approach to AI bias," a trend in AI literature where biases are seen as separate occurrence linked to specific stages in an AI pipeline. Dealing with these multiple biases can trigger a sense of excessive overload in managing each potential bias individually or promote the (...) adoption of an uncritical approach to understanding the influence of biases in developers’ decision-making. The BNA fosters dialogue and a critical stance among developers, guided by external experts, using graphical representations to depict biased connections. To test the BNA, we conducted a pilot case study on the "waiting list” project, involving a small AI developer team creating a healthcare waiting list NPL model in Chile. The analysis showed promising findings: (i) the BNA aids in visualizing interconnected biases and their impacts, facilitating ethical reflection in a more accessible way; (ii) it promotes transparency in decision-making throughout AI development; and (iii) more focus is necessary on professional biases and material limitations as sources of bias in AI development. (shrink)

Have you been surprised by the recent development and diffusion of generative artificial intelligence (AI)? Many institutions of civil society have been caught off guard, which provides them with motivation to think ahead. And as many new plausible pathways of socio-technical development are opening up, a growing interest in technology ethics that addresses our corresponding moral uncertainties is warranted. In Sven Nyholm’s words, “[t]he field of technology ethics is absolutely exploding at the moment” (p. 262), and so the publication of (...) his introduction to the topic is well timed. (shrink)

Through the lens of feminist speech act theory, this paper argues that artificial intelligence romance systems objectify and subordinate nonvirtual women. AI romance systems treat their users as consumers, offering them relational invulnerability and control over their (usually feminized) digital romantic partner. This paper argues that, though the output of AI chatbots may not generally constitute speech, the framework offered by an AI romance system communicates an unjust perspective on intimate relationships. Through normalizing controlling one’s intimate partner, these systems operate (...) as speakers that rank women as non-player characters and license their oppression, which is unjust even in the absence of empirical harms. Understanding AI romance systems as speakers has implications for policymakers. First, this account helps close the so-called responsibility gap in the operation of an AI system. Second, it places these systems under the purview of institutions’ policies regarding hate speech, suggesting one avenue for arguing that these systems may violate extant policies. Finally, it provides support for the pursuit of empirical research into AI romance systems’ effects on users. (shrink)

Training artificial intelligence (AI) systems requires vast quantities of data, and AI developers face a variety of barriers to accessing the information they need. Synthetic data has captured researchers’ and industry’s imagination as a potential solution to this problem. While some of the enthusiasm for synthetic data may be warranted, in this short paper we offer critical counterweight to simplistic narratives that position synthetic data as a cost-free solution to every data-access challenge—provocations highlighting ethical, political, and governance issues the use (...) of synthetic data can create. We question the idea that synthetic data, by its nature, is exempt from privacy and related ethical concerns. We caution that framing synthetic data in binary opposition to “real” measurement data could subtly shift the normative standards to which data collectors and processors are held. And we argue that by promising to divorce data from its constituents—the people it represents and impacts—synthetic data could create new obstacles to democratic data governance. (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)

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)

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)

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.

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)

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)

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)

The demands of fair machine learning are often expressed in probabilistic terms. Yet, most of the systems of concern are deterministic in the sense that whether a given subject will receive a given score on the basis of their traits is, for all intents and purposes, either zero or one. What, then, can justify this probabilistic talk? We argue that the statistical reference classes used in fairness measures can be understood as defining the probability that hypothetical persons, who are representative (...) of social roles, will receive certain goods. We call these hypothetical persons “representative individuals.” We claim that what we owe to actual, concrete individuals—whose individual chances of receiving the good in the system might be extreme (i.e., either zero or one)—is that their representative individual has an appropriate probability of receiving the good in question. While less immediately intuitive than other approaches, we argue that the representative individual approach has important advantages over other ways of making sense of this probabilistic talk in the context of fair machine learning. (shrink)

There is a theorem that shows that it is impossible for an algorithm to jointly satisfy the statistical fairness criteria of Calibration and Equalized Odds non-trivially. But what about the recently advocated alternative to Calibration, Base Rate Tracking? Here we show that Base Rate Tracking is strictly weaker than Calibration, and then take up the question of whether it is possible to jointly satisfy Base Rate Tracking and Equalized Odds in non-trivial scenarios. We show that it is not, thereby establishing (...) an even more general impossibility theorem. (shrink)

As artificial intelligence (AI) systems increasingly permeate processes of cultural and epistemic production, there are growing concerns about how their outputs may confine individuals and groups to static or restricted narratives about who or what they could be. In this paper, we advance the discourse surrounding these concerns by making three contributions. First, we introduce the concept of aspirational affordance to describe how technologies of representation---paintings, literature, photographs, films, or video games---shape the exercising of imagination, particularly as it pertains to (...) possibilities for agency. Second, we provide three reasons for scrutinizing of AI's influence on aspirational affordances: AI's influence is potentially more potent, but less public than traditional sources; AI's influence is not simply incremental, but ecological, transforming the entire landscape of cultural and epistemic practices that traditionally shaped aspirational affordances; and AI's influence is highly concentrated, with a few corporate-controlled systems mediating a growing portion of aspirational possibilities. Third, to advance such a scrutiny of AI's influence, we introduce the concept of aspirational harm, which arises when AI-generated affordances distort or diminish a group's hermeneutical resources for imagining their practical possibilities. Through three case studies, we illustrate how aspirational harms extend the existing discourse on AI-inflicted harms beyond representational and allocative harms, warranting separate attention. Through these conceptual resources and analyses, this paper advances understanding of the psychological and societal stakes of AI’s role in shaping individual and collective aspirations. (shrink)

Is the pool of fair predictive algorithms fair? It depends, naturally, on both the criteria of fairness and on how we pool. We catalog the relevant facts for some of the most prominent statistical criteria of algorithmic fairness and the dominant approaches to pooling forecasts: linear, geometric, and multiplicative. Only linear pooling, a format at the heart of ensemble methods, preserves any of the central criteria we consider. Drawing on work in the social sciences and social epistemology on the theoretical (...) foundations of the wisdom of crowds, we explain how our observations present an exception to the general trend of finding tradeoffs between the accuracy and fairness of forecasts. (shrink)

When selecting a model from a set of equally performant models, how much unfairness can you really reduce? Is it important to be intentional about fairness when choosing among this set, or is arbitrarily choosing among the set of “good” models good enough? Recent work has highlighted that the phenomenon of model multiplicity—where multiple models with nearly identical predictive accuracy exist for the same task—has both positive and negative implications for fairness, from strengthening the enforcement of civil rights law in (...) AI systems to showcasing arbitrariness in AI decision-making. Despite the enormous implications of model multiplicity, there is little work that explores the properties of sets of equally accurate models, or Rashomon sets, in general. In this paper, we present five main theoretical and methodological contributions which help us to understand the relatively unexplored properties of the Rashomon set, in particular with regards to fairness. Our contributions include methods for efficiently sampling models from this set and techniques for identifying the fairest models according to key fairness metrics such as statistical parity. We also derive the probability that an individual’s prediction will be flipped within the Rashomon set, as well as expressions for the set’s size and the distribution of error tolerance used across models. These results lead to policy-relevant takeaways, such as the importance of intentionally looking for fair models within the Rashomon set, and understanding which individuals or groups may be more susceptible to arbitrary decisions. (shrink)

The literature on algorithmic fairness has examined exogenous sources of biases such as shortcomings in the data and structural injustices in society. It has also examined internal sources of bias as evidenced by a number of impossibility theorems showing that no algorithm can concurrently satisfy multiple criteria of fairness. This paper contributes to the literature stemming from the impossibility theorems by examining how informational richness affects the accuracy and fairness of predictive algorithms. With the aid of a computer simulation, we (...) show that informational richness is the engine that drives improvements in the performance of a predictive algorithm, in terms of both accuracy and fairness. The centrality of informational richness suggests that classification parity, a popular criterion of algorithmic fairness, should be given relatively little weight. But we caution that the centrality of informational richness should be taken with a grain of salt in light of practical limitations, in particular, the so-called bias-variance trade off. (shrink)

Statistical fairness criteria are widely used for diagnosing and ameliorating algorithmic bias. However, these fairness criteria are controversial as their use raises several difficult questions. I argue that the major problems for statistical algorithmic fairness criteria stem from an incorrect understanding of their nature. These criteria are primarily used for two purposes: first, evaluating AI systems for bias, and second constraining machine learning optimization problems in order to ameliorate such bias. The first purpose typically involves treating each criterion as a (...) necessary condition for fairness. The second use involves treating criteria as sufficient conditions for fairness. Since the criteria are used for both roles, some researchers have treated them as both necessary and sufficient conditions, i.e., as definitions of algorithmic fairness. However, serious problems have been raised for the use of these fairness criteria. Under ordinary circumstances, it is impossible to satisfy multiple criteria at the same time. Moreover, there are counterexamples to both the sufficiency and necessity for fairness of each criterion. I argue that we should instead understand fairness criteria as merely providing evidence of fairness. In other words, satisfaction (or violation) of these criteria should be understood as potential evidence of fairness (or bias). Whether a criterion counts as evidence in a particular case will depend on stakeholders' background knowledge and the specific features of the system's task. This evidence account of fairness conditions provides guidance for recognizing both the appropriate uses and the limitations of fairness criteria. (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)