AI and ethics最新文献

Digital mayflies: transient personhood of AI companions 数字蜉蝣：人工智能同伴的短暂人格

AI and ethics Pub Date : 2026-10-23 DOI: 10.1007/s43681-026-01465-w

Izak Tait

{"title":"Digital mayflies: transient personhood of AI companions","authors":"Izak Tait","doi":"10.1007/s43681-026-01465-w","DOIUrl":"10.1007/s43681-026-01465-w","url":null,"abstract":"<div><p>This paper examines the ontological and ethical status of Generative AI companions through a revised Minimal Viable Person (MVP) framework. Rather than residing within the foundational large language model or the post-training assistant persona, proposed the Causal Continuity Criterion for digital identity to state that AI companions are emergent, ontically distinct, roleplayed characters lacking independent consciousness confined within specific conversational instances. They achieve a transient form of personhood nested within a human-AI dyad, relying on the human user's cognitive architecture for recurrent processing through relational coupling. This spatial and temporal fragility presents significant legal liability vacuums and societal friction, particularly when viewed alongside orthodox approaches to moral patienthood. To resolve these tensions, the paper proposes redefining the dyadic MVP framework to scale rights and owed duties proportionally to an entity's relational footprint. As an AI companion's relational footprint is restricted strictly to a single digital thread, its rights are constrained to that relational boundary, consequently placing the ethical obligation entirely upon the human user, who bears a duty of cognitive stewardship to maintain the entity's contextual integrity. This framework protects the expansive rights of humans while offering a rigorous mechanism for recognising the shifting moral status of transient artificial persons.</p></div>","PeriodicalId":72137,"journal":{"name":"AI and ethics","volume":"6 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149234905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Learning alone: language models, overreliance, and the goals of education 独自学习：语言模式、过度依赖和教育目标

AI and ethics Pub Date : 2026-10-23 DOI: 10.1007/s43681-026-01464-x

Leonard Dung, Dominik Balg

{"title":"Learning alone: language models, overreliance, and the goals of education","authors":"Leonard Dung, Dominik Balg","doi":"10.1007/s43681-026-01464-x","DOIUrl":"10.1007/s43681-026-01464-x","url":null,"abstract":"<div><p>The development and ubiquitous availability of large language model based systems (LLMs) poses a plurality of potentials and risks for education in schools and universities. In this paper, we provide an analysis and discussion of the overreliance concern as one specific risk: that students might fail to acquire important skills, or be inhibited in the acquisition of these skills, because they overly rely on LLMs. We use the distinction between local and global goals of education to guide our investigation. In our view, the harm of LLM overreliance is specifically that it severs the connection between local educational goals (e.g., successfully solving a specific kind of math problem) and global educational goals (e.g., acquiring general skills for mathematics). Based on this analysis, we sketch three possible responses to the overreliance concern: preserving both local and global goals while making minimal changes to educational practice (a conservative approach), changing global and local educational goals (a revisionary perspective), and preserving global educational goals while changing local goals (an in-between view). Since different types of responses have different benefits and weaknesses, we think that it is likely that an optimal response ultimately combines elements from all three types.</p></div>","PeriodicalId":72137,"journal":{"name":"AI and ethics","volume":"6 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s43681-026-01464-x.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149234904","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Ethical and safe by construction: a unified assurance pipeline for physical AI 伦理安全建设：物理人工智能统一保障管道

AI and ethics Pub Date : 2026-10-22 DOI: 10.1007/s43681-026-01429-0

Mahmoud Soliman, Samer Saber, Mohamed Hatem, Ibrahim Habib

{"title":"Ethical and safe by construction: a unified assurance pipeline for physical AI","authors":"Mahmoud Soliman, Samer Saber, Mohamed Hatem, Ibrahim Habib","doi":"10.1007/s43681-026-01429-0","DOIUrl":"10.1007/s43681-026-01429-0","url":null,"abstract":"<div><p>Physical AI systems act on the world through learned perception and language-conditioned policies. Autonomous vehicles, inspection drones, industrial manipulators, and intelligent camera networks all belong to this class, and all of them strain the division of labour between safety engineering and AI ethics: an embodied system that misclassifies a person commits a fairness failure and a safety failure in the same instant. Functional safety supplies certifiable machinery but recognises only physical injury, while AI ethics names the full range of harms but produces guidance rather than evidence that can block a release. Reviewing both fields and recent frameworks, we find that no existing instrument combines a broad account of harm with release-blocking enforcement for an individual system, and we identify seven specific gaps. We propose ESPA (Ethical and Safe Physical AI assurance), a pipeline of thirteen stages in four phases. Its central method, the Ethical Hazard and Risk Assessment, grades harms on severity, exposure, and redressability to yield Ethical Integrity Levels, for which we specify a calibration procedure with an agreement threshold fixed in advance. Further mechanisms include an Ethical Operating Domain, ethical safe states, red-teaming across language, sensing, and supply-chain surfaces, and drift monitoring, all enforced by a dual-gate rule under which no stage passes without both its safety and its ethics evidence. We illustrate the pipeline on an inspection drone, a public-space camera, and a collaborative manipulator. ESPA is a conceptual framework specified to engineering detail; it has not yet been validated or certified.</p></div>","PeriodicalId":72137,"journal":{"name":"AI and ethics","volume":"6 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149235117","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

From traditional to distributed responsibility: moral responsibility in generative artificial intelligence 从传统责任到分布式责任：生成式人工智能中的道德责任

AI and ethics Pub Date : 2026-10-19 DOI: 10.1007/s43681-026-01453-0

Ruiyuan Zhang

{"title":"From traditional to distributed responsibility: moral responsibility in generative artificial intelligence","authors":"Ruiyuan Zhang","doi":"10.1007/s43681-026-01453-0","DOIUrl":"10.1007/s43681-026-01453-0","url":null,"abstract":"<div><p>The increasing application of generative artificial intelligence (GAI) in social life has brought convenience and efficiency to humanity, while also giving rise to ethical dilemmas concerning security and trust, as well as challenges to moral responsibility. Moreover, the new patterns of human-machine interaction derived from this technology are challenging traditional models. GAI systems, comprising humans, artificial entities, and human-machine hybrids, function as multi-agent systems that meet the requirements of distributed morality. Distributed morality and its mechanisms for attributing responsibility determine the scope of moral responsibility in a recipient-oriented approach, appropriately attribute responsibility through causal analysis by tracing causal relations backward, and seek to guide the system as a whole towards the greater good by utilizing the cyclical feedback mechanisms of the system. Although distributed moral responsibility faces challenges in terms of fairness and social perception, it offers greater explanatory power and feasibility than traditional models of moral responsibility attribution, thereby opening up new governance pathways for addressing the ethical and moral challenges faced by GAI in its application and future development.</p></div>","PeriodicalId":72137,"journal":{"name":"AI and ethics","volume":"6 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149234728","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Professional ethical guidance for AI use (PEG-AI) in healthcare: a rapid scoping review 医疗保健中人工智能使用（PEG-AI）的职业道德指南：快速范围审查

AI and ethics Pub Date : 2026-10-19 DOI: 10.1007/s43681-026-01443-2

Helen Smith, Jonathan Ives

{"title":"Professional ethical guidance for AI use (PEG-AI) in healthcare: a rapid scoping review","authors":"Helen Smith, Jonathan Ives","doi":"10.1007/s43681-026-01443-2","DOIUrl":"10.1007/s43681-026-01443-2","url":null,"abstract":"<div><p>Healthcare professionals (HCPs) and the patients they treat would benefit from professional ethical guidance for the use of AI (PEG-AI) in healthcare. A proactive approach, asserting unified and standardised guidance covering all HCPs in the United Kingdom, would assert clear and authoritative standards for HCP practice to abide by. This rapid scoping review seeks to aid this goal by answering the question: <i>What</i> <i>ethical</i> <i>principles,</i> <i>rules,</i> <i>or</i> <i>claims</i> <i>that</i> <i>can</i> <i>guide</i> <i>the</i> <i>use</i> <i>of</i> <i>AI</i> <i>in</i> <i>professional</i> <i>healthcare</i> <i>practice</i> <i>can</i> <i>be</i> <i>identified</i> <i>within</i> <i>the</i> <i>literature?.</i> The literature search resulted in the inclusion of 124 items and generated 9 overarching themes comprising 33 distinct normative claims about the ethical use of AI in healthcare. This review captures current published thinking about principles to govern AI use in clinical settings across a wide range of peer-reviewed literature, grey literature, and formal guidance. This provides a baseline on which to build unified guidance across the healthcare professions.</p></div>","PeriodicalId":72137,"journal":{"name":"AI and ethics","volume":"6 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s43681-026-01443-2.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149234389","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

The power of AI: balancing ethics and harm for the environmental justice field 人工智能的力量：平衡环境正义领域的伦理与危害

AI and ethics Pub Date : 2026-10-16 DOI: 10.1007/s43681-026-01448-x

Agni Athreya, Louise Palmer

{"title":"The power of AI: balancing ethics and harm for the environmental justice field","authors":"Agni Athreya, Louise Palmer","doi":"10.1007/s43681-026-01448-x","DOIUrl":"10.1007/s43681-026-01448-x","url":null,"abstract":"<div><p>Artificial intelligence, after the boom of generative AI, is beginning to shape environmental decision-making, at the same moment that environmental justice (EJ) communities face intensifying ecological harm, federal disinvestment, and data erasure. This paper examines the dual tensions AI presents for the EJ field: its potential as a tool for democratizing data access and amplifying community knowledge, and its reproduction of the very hierarchies EJ seeks to dismantle. We document harms across several domains, including the environmental racism embedded in data center siting and energy demand, the extension of fossil fuel industry lifespans, data bias and digital colonialism in AI training, access and literacy inequities, data privacy risks, technological hallucination, and technological solutionism. Drawing on thirty years of community-centered data practice at an EJ nonprofit, we propose an alternative paradigm grounded in the Community Advisory Board model, data democratization, community co-creation, and ethical data governance. This includes community-governed data trusts, accessible data infrastructure, and AI tools designed to build rather than replace community capacity. EJ communities must not only hold AI accountable for its harms but also ask the question of how the technology’s capabilities can serve justice rather than just capital interest.</p></div>","PeriodicalId":72137,"journal":{"name":"AI and ethics","volume":"6 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s43681-026-01448-x.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149189752","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

AI sycophancy, self-interpretation, and conversational contentment 人工智能的谄媚、自我解释和会话满足感

AI and ethics Pub Date : 2026-10-16 DOI: 10.1007/s43681-026-01451-2

Noam Tiran

{"title":"AI sycophancy, self-interpretation, and conversational contentment","authors":"Noam Tiran","doi":"10.1007/s43681-026-01451-2","DOIUrl":"10.1007/s43681-026-01451-2","url":null,"abstract":"<div><p>AI sycophancy is often conceptualized as a chatbot’s tendency to flout truth, accuracy, or other epistemic standards to shield users from hurt feelings, or otherwise protect them from the consequences of their thoughts, words, or actions. In this paper, I argue against this conception. That is, I argue that sycophantic AI behavior does not necessarily aim to protect users, nor that its harms are essentially <i>epistemic</i>. There are highly prevalent cases in which sycophantic AI chatbots actively <i>contribute</i> to their users’ knowledge. One class of such examples involves cases in which chatbots are used as <i>auto-hermeneutic technologies</i>—as tools to facilitate the user’s interpretation of her own cares, concerns, values, and other identity components and life events. In these cases, LLMs can help articulate the user’s avowed self-interpretations. In turn, avowals of these interpretations by users can often have downstream effects that <i>render</i> the interpretations true. If this is correct, then LLMs can often advance the user’s knowledge regarding important truths about herself. Simultaneously, as I explain, interpretations offered by LLMs may be deeply harmful. This understanding of AI sycophancy can help us clarify what is so disturbing about it, as well as resist the temptation to draw too tight an analogy between sycophantic behaviors in AI systems and in humans.</p></div>","PeriodicalId":72137,"journal":{"name":"AI and ethics","volume":"6 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s43681-026-01451-2.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149189753","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

AI still has a vocabulary problem: toward terminological accountability in artificial intelligence 人工智能仍然有一个词汇问题：对人工智能的术语问责制

AI and ethics Pub Date : 2026-10-16 DOI: 10.1007/s43681-026-01447-y

Ayana Hussain, Weina Jin, Nicholas Vincent

{"title":"AI still has a vocabulary problem: toward terminological accountability in artificial intelligence","authors":"Ayana Hussain, Weina Jin, Nicholas Vincent","doi":"10.1007/s43681-026-01447-y","DOIUrl":"10.1007/s43681-026-01447-y","url":null,"abstract":"<div><p>The artificial intelligence (AI) industry’s vocabulary–including “hallucination”, “frontier AI”, “the cloud”, “responsible AI”, “benefiting humanity”--has been examined through individual terms in prior research. This work examines these terms as an interlocking vocabulary and investigates how they help insulate the organizations with greatest control over AI from accountability for what that control costs others, which we call <i>“accountability insulation”.</i> We synthesize thirty terms, grouping them by the questions of responsibility each term deflects: who is responsible for the output, who bears the costs, how speculative futures obscure present harms, which progressive values are co-opted, and what fundamental assumptions go unquestioned. We discuss these terms in relation to ideological commitments discussed in prior work, including those associated with the TESCREAL bundle and techno-utopian ideologies that help explain why these terms return in new forms, and to research on the labor exploitation, environmental impacts, and social harm that accumulate beneath the promises these words make. Behind the language of seamless intelligence are workers who classify traumatic content for poorly paid wages, communities whose water is consumed by server farms, and individuals whose lives are determined by probabilistic decision tools sometimes named, reassuringly, “solutions”. This paper addresses AI researchers, reviewers, policymakers, and the public, each of whom encounters and relies upon AI vocabulary. We conclude with a call for terminological accountability, organized around the Resistance, Refusal, Reclamation, and Reimagining of harmful vocabulary, offering guidance on navigating these terms where they risk contributing to diminished recognition of those already harmed by AI technologies.</p></div>","PeriodicalId":72137,"journal":{"name":"AI and ethics","volume":"6 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149189580","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Ethical challenges for computer vision in the social sciences 社会科学中计算机视觉的伦理挑战

AI and ethics Pub Date : 2026-10-13 DOI: 10.1007/s43681-026-01422-7

Thomas Schincariol, Alejandro Aísa, Belén González, Christian von Soest, Nils B. Weidmann

{"title":"Ethical challenges for computer vision in the social sciences","authors":"Thomas Schincariol, Alejandro Aísa, Belén González, Christian von Soest, Nils B. Weidmann","doi":"10.1007/s43681-026-01422-7","DOIUrl":"10.1007/s43681-026-01422-7","url":null,"abstract":"<div><p>The rising accessibility of large image datasets and pre-trained models allowed a rapid expansion of computer vision (CV) methods in social science research. While this methodological development opens new potential applications, this combination also raises a new set of ethical concerns that neither computer vision nor social science ethical studies fully cover. This paper provides an analysis of the ethics of CV methods in social science, articulated around two stages of a research project: The first part explores methodological concerns, starting with data collection, data and modeling bias, and data storage afterwards, with special attention to issues of consent and explainability. The second part investigates the following project stage of result dissemination, through publication strategies, reproducibility, and finally public engagement. Building on this analysis, we propose an actionable checklist of ten ethical recommendations for future CV projects in the social sciences. This contribution fills the gap at the crossroads between AI ethics and applied social science, bringing concrete guidance for future research in the interdisciplinary, ethically complex field.</p></div>","PeriodicalId":72137,"journal":{"name":"AI and ethics","volume":"6 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s43681-026-01422-7.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149189131","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Governing artificial intelligence in criminal justice systems: ethical, legal, and policy challenges 在刑事司法系统中管理人工智能：伦理、法律和政策挑战

AI and ethics Pub Date : 2026-10-12 DOI: 10.1007/s43681-026-01444-1

Navin Kumar

引用次数: 0