AI and ethics最新文献

Artificial intelligence (AI)’s rapid advancement poses opportunities and challenges for environmental ethics. While AI has the potential to enhance ecological sustainability through data-driven solutions, it also risks depersonalizing ethical decision-making and reinforcing a technocratic paradigm that prioritizes efficiency over human dignity and environmental stewardship. This paper explores how Karol Wojtyła’s personalist philosophy provides a sound ethical framework for addressing these concerns. Personalism, which emphasizes the irreducibility of the human person, responsibility, and relationality, offers a foundation for rethinking environmental ethics in the AI era. This study supports a person-centered approach to environmental decision-making by including Wojtyła’s ideas on human agency, the common good, and ecological responsibility. Such an approach resists the reduction of human moral agency to algorithmic processes while fostering solidarity, subsidiarity, and ecological justice. Ultimately, this paper argues that a personalist environmental ethic can guide technological development toward serving humanity and the natural world, ensuring that AI remains a tool for sustainable and ethical progress rather than an autonomous arbiter of ecological fate.

The rapid advancement of artificial intelligence (AI) has introduced profound societal and ethical challenges, necessitating a paradigm shift in AI system design. This paper introduces a novel framework that enables AI systems to design, audit, and evolve themselves ethically, through an adaptation of the echeloned design science research (eDSR) methodology. These AI systems will certainly evolve beyond mere tools to design, refine, and govern themselves within ethical constraints. The framework embeds four core principles: responsible autonomy, where AI systems self-regulate their decisions within ethical boundaries; AI self-explainability, enabling AI-to-AI transparency and internal decision auditing; AI bootstrapping, supporting iterative self-enhancement; and knowledge-informed machine learning (KIML), which integrates domain expertise for context-aware learning. We extend the concept of AI-as-a-User-of-AI, wherein autonomous AI agents behave as collaborative entities that engage in structured dialogues to refine decisions and enforce ethical alignment. Unlike traditional systems that rely on human-in-the-loop oversight or post-hoc explanations, our framework allows AI to monitor and evolve its reasoning in real time. By embedding ethical reasoning, self-explanation, and learning directly into system architecture through modular design echelons, the proposed generative eDSR (GeDSR) framework combines eDSR’s structured and multi-phased approach with AI-to-AI collaboration, which enables scalability, adaptability, and ethical alignment across diverse applications. By embedding ethical reasoning and iterative learning at the architectural level, the proposed framework promotes the development of self-improving AI systems aligned with human values, thus laying the groundwork for a shift from human-dependent oversight to a resilient, AI-centric ecosystem.

The integration of artificial intelligence (AI) into human resource development (HRD) demands a re-examination of learning theories, developmental practices, and ethical frameworks. This integrative review synthesizes sociotechnical systems theory, adult learning models, augmentation strategies, technology adoption frameworks, and ethics considerations to build a comprehensive conceptual model of AI-enhanced HRD. Key themes include the joint optimization of human and machine capabilities, the personalization and critical interpretation of AI-mediated learning, and the proactive stewardship of fairness, transparency, and inclusion. Drawing upon contemporary studies (2021–2025) and emerging empirical evidence, this paper offers an integrative model that links technical, social, and ethical subsystems within HRD. Cross-cultural dimensions and methodological innovations for studying AI-HRD dynamics are also discussed. The paper offers theoretical contributions to HRD scholarship and practical recommendations for designing adaptive, ethical, and human-centered learning ecosystems in the age of intelligent technologies. Ultimately, the model situates HRD as an active agent in shaping responsible AI futures that enhance, rather than erode, human learning and development.

This article proposes the Universal Theory of Core Values in Intelligent Systems (UTCVIS), a framework for understanding how value-like constraints can emerge within intelligent agents as a function of structural pressures rather than external moral imposition. UTCVIS treats values not as intrinsically human mental states, but as stability-promoting heuristics that arise when agents seek continuity in open, resource-coupled environments with repeated interaction and externalities. Drawing on systems theory, evolutionary biology, institutional economics, and multi-agent reinforcement learning, the framework argues that patterns functionally analogous to honesty, fairness, reciprocity, and stewardship can emerge when cooperation, information integrity, and credible sanctioning are necessary for long-term viability. Rather than framing AI ethics solely as a problem of encoding human values into machines, UTCVIS reframes alignment as a coexistence problem: under what environmental and institutional conditions will the survival logics of advanced AI systems remain compatible with human flourishing? The article articulates six foundational principles spanning energy maintenance, continuity, environmental pressure, social complexity, core values, and ethical reflection, and links each principle to observable system properties suitable for empirical investigation. By situating these claims within debates on the is–ought gap, moral universalism versus relativism, anthropomorphism, and AI control, UTCVIS offers a structured account of emergent values that is both philosophically grounded and empirically generative. The paper concludes by outlining simulation pathways and limitations, emphasizing caution in interpreting emergent “values” in non-biological systems while arguing for their relevance to AI ethics and governance.

Medical devices are required to meet a variety of stringent regulations in markets around the world. The EU recently introduced AI legislation, and the UK government has issued AI regulatory guidance that is cascading to medical regulators. The introduction of generative AI tools into medical device development requires a detailed understanding of how these new and existing regulations may interact, as well as the underpinning ethical risks, and yet information in this area is scarce. A hypothetical medical device related use-case was created to highlight risks. A product development process was explored to elucidate the impacts of generative AI inputs and outputs. Generative AI risks are varied and prevalent across most areas of medical device businesses, particularly where traceability and reproducibility of information is key. These risks were consolidated into a UK focussed ethical framework that considered business, employee, customer, and regulator needs. The distinct approaches of different regions to generative AI regulation create challenges for businesses and regulators, which may create confusion or delays for those seeking to integrate the technology into fields with strict extant legal requirements. Simultaneously, the pace of generative AI adoption is relentless. An ethical framework that considers the key tenets of both nascent AI and established medical device guidance and regulations is necessary to protect medical device businesses and avoid significant duplication of regulatory effort. Such a framework also aids in anticipating potential future regulatory developments.

Principlism is one of the leading approaches to AI ethics. Developed originally to address bioethical issues in medicine and research, the theory requires decision-makers to consider various ethical principles to justify their actions. Despite its dominance in bioethics and growing popularity in AI ethics, principlism faces serious theoretical and practical criticisms. One serious criticism is that since ethical principles are combined ad hoc, conflicts inevitably arise between them, leading to inconsistencies. While defenders of principlism propose a process for resolving such disputes, contemporary critics argue that this process is incomplete, and at best, principlist frameworks can only help structure analysis and justifications intelligently, but cannot provide definitive, action-guiding moral prescriptions. AI principlists have not adequately reckoned with this theoretical limitation. Here, I propose a solution to conflicts between principles by designating one principle as an arbitrating principle above others—what I call hierarchical principlism. Since attempts to use existing principles as arbiters have led to controversy, I suggest using a new principle—a modified version of the principle of beneficence requiring the minimization of suffering, which I call the principle of compassion—to arbitrate these conflicts. I argue that this approach, which I call compassionate principlism, leads to fewer moral objections and inconsistencies and provides more definitive action-guiding moral prescriptions in AI ethics than traditional principlism. I conclude by applying compassionate principlism to ethical dilemmas in AI ethics, including misinformation, bias, and automation.

In this essay, we explore the parts of ethical life that cannot be captured and may actually be foreclosed by the dominant principle-centered approach to AI ethics. We argue that principles maintain traces of the moral encounters they were abstracted from, including traces of the moral and metaphysical intuitions that guided people in those encounters. Principlism is thus haunted by an unacknowledged particularity and what we call a “spectral moral ontology.” While bioethics principlism remains connected to the animating force of the encounter through a moral actor, who vivifies its spectral moral ontology, AI possibly removes this moral actor. To do so, programmers must further abstract from the moral encounter by reducing the principles to formal logics cutting “ethics” off from its animating force. This approach to AI ethics creates and deploys a reductive spectral moral ontology that is woefully inadequate to the complexity and irreducibility of the moral encounter.

Artificial intelligence (AI) in healthcare has led to many promising developments; however, increasingly, AI research is funded by the private sector leading to potential trade-offs between benefits to patients and benefits to industry. Health AI practitioners should prioritize successful adaptation into clinical practice in order to provide meaningful benefits to patients, but translation usually requires collaboration with industry. We discuss three features of AI studies that hamper the integration of AI into clinical practice from the perspective of researchers and clinicians. These include lack of clinically relevant metrics, lack of clinical trials and longitudinal studies to validate results, and lack of patient and physician involvement in the development process. For partnerships between industry and health research to be sustainable, a balance must be established between patient and industry benefit. We propose three approaches for addressing this gap: improved transparency and explainability of AI models, fostering relationships with industry partners that have a reputation for centering patient benefit in their practices, and prioritization of overall healthcare benefits. With these priorities, we can sooner realize meaningful AI technologies used by clinicians where mutually beneficial impacts for patients, healthcare providers, and industry can be realized.

As more complex Artificial Intelligence Systems (AISs) become increasingly embedded in critical sectors, the “black box dilemma” has emerged as a key concern in both technical and legal debates. Unfortunately, the concepts most frequently invoked to formulate and address the epistemic dimension of AISs—transparency, traceability, interpretability, and explainability—remain ill-defined and inconsistently applied in EU regulation. This article proposes a regulatory taxonomy that distinguishes these four concepts as layered and interdependent dimensions of AI opacity, each with distinct epistemic and normative roles. While each of these concepts offers a necessary but partial view of AI opacity, none is sufficient on its own. They support a complete understanding of AIS outcomes only when considered together, as interdependent but connected layers. Thus, the taxonomy provides a conceptual framework for legal interpretation, compliance strategies, and informed future legislative design. The article illustrates the new framework through a case study on algorithmic credit scoring. By clarifying the distinct functions and audiences of each concept, the article contributes to a more coherent regulatory approach to AI opacity, one that enables accountability, fosters innovation, and strengthens trust in automated decision-making.

This article examines the legal and ethical responsibilities of data controllers when using artificial intelligence (AI) to process personal data, with a focus on safeguarding vulnerable individuals. It develops a structured, risk-based responsibility framework grounded in European data protection law, particularly the General Data Protection Regulation (GDPR), the Law Enforcement Directive (LED), and the emerging obligations of the European Union’s (EU) AI Act. The framework seeks to reconcile technological innovation with the duty to uphold fundamental rights. Using a doctrinal legal research method, the study analyses legislation, case law, and regulatory guidance to identify three core areas of responsibility: risk assessment and mitigation, proportionality in protective measures, and demonstrable compliance. Procedural tools such as Data Protection Impact Assessments (DPIAs), Fundamental Rights Impact Assessments (FRIAs), data protection by design and default, and human oversight are integrated into this model. The framework’s practical relevance is illustrated through supervisory authority decisions and public debates in education, employment, credit scoring, and smart-city surveillance. The findings show that AI technologies—especially profiling, biometric recognition, and automated decision-making—create heightened risks for rights, including privacy, dignity, autonomy, and non-discrimination. Effective governance requires continuous risk evaluation, safeguards proportionate to the context, and, in highand persistent residual risk cases, consultation with supervisory authorities. The article concludes that protecting vulnerable individuals in AI contexts requires a proactive and adaptive governance model. While challenges and jurisdictional limits remain, the proposed framework offers a legally sound and ethically grounded basis for responsible AI deployment that prioritizes those most at risk.