Abstract There are underappreciated moral costs for deploying artificially intelligent agents in our present bureaucratically and market‐structured world. Currently, AI systems lack the interiority and mutual vulnerability required for genuine moral relationality. As such, their integration into social “nodes,” or places of individual and organizational interaction, make our moral ecology more austere, reducing morally sensitive interactions. In turn, this loss of morally sensitive nodes reduces the opportunities for developing moral skills, which typically benefit from moral and dialogic friction. The result is increased pressure to defer to institutional defaults and to tolerate an erosion of individual moral agency. Finally, this situation may decrease our desire for mutuality. Each of these things has consequences for Christian understandings of imago Dei , the epistemology of ethics, and ethical mutuality.
Moral reasoning is fundamental to safe Artificial Intelligence (AI), yet ensuring its consistency across modalities becomes critical as AI systems evolve from text-based assistants to embodied agents. Current safety techniques demonstrate success in textual contexts, but concerns remain about generalization to visual inputs. Existing moral evaluation benchmarks rely on textonly formats and lack systematic control over variables that influence moral decision-making. Here we show that visual inputs fundamentally alter moral decision-making in state-of-the-art (SOTA) Vision-Language Models (VLMs), bypassing text-based safety mechanisms. We introduce Moral Dilemma Simulation (MDS), a multimodal benchmark grounded in Moral Foundation Theory (MFT) that enables mechanistic analysis through orthogonal manipulation of visual and contextual variables. The evaluation reveals that the vision modality activates intuition-like pathways that override the more deliberate and safer reasoning patterns observed in text-only contexts. These findings expose critical fragilities where language-tuned safety filters fail to constrain visual processing, demonstrating the urgent need for multimodal safety alignment.
According to the Doctors of the Church, God is a divine Lawgiver who may give a command that overrides an absolute moral norm. In this article, some of the modern and ancient scenarios in which absolute moral norms are superseded by divine authority will be explored and discussed. This article will also consider Francis’s description of how conscience can generate such “exceptions” through a process of personal and pastoral discernment in the concrete lives of specific persons. In conclusion, the article considers the Magisterium’s articulation of how moral norms and conscience are not destined to contradict each other but rather converge as mutual and complementary representatives of divine authority.
Artificial intelligence (AI) is a fundamental breakthrough in contemporary science. At its innovative core is the quest to simulate human intelligence processes using machines and computer systems. A topical moral discussion lies in the development of sexual robots that can perform conjugal functions. However, the Bible presents the concept of sexuality as a privilege exercised within the precincts of the same species in heterosexual relationships. Hence, according to the Scriptures, human beings can only have a sexual relationship with fellow human beings of the opposite biological gender. However, the rise in robotic technology that includes sexuality raises fundamental questions. What are the ethical implications of AI sexual innovation for marriage covenants? Towards what cause should contemporary theology and ethics relate to this innovation within AI? What is the place of sexuality in humans? Can sexuality with a human-like robot be understood as a biblical and legitimate alternative in the face of rising sexually transmitted diseases? This theoretical paper seeks to interrogate this line of AI innovation from an ethical and theological assessment.
This article aims to reconstruct the meaning of fasād (enviromental damage) in the Qur’an through a tafsir maqāṣidīapproach, serving as a theological response to the issue of global warming. Contemporary phenomena of environmental degradation, such as pollution, deforestation, and extreme climate change, are understood as tangible manifestations of the concept of fasād, as referenced in Q.S. Ar-Rūm: 41. This study employs qualitative research methods, utilizing library sources and the maqāṣidī interpretative approach as its analytical framework. The findings indicate that the understanding of ecological verses has evolved from classical textual interpretations to scientific and logical approaches during the medieval period, and further to contextual interpretations that address contemporary environmental challenges. The reconstruction of the meaning of fasād is highly relevant in developing an Islamic ecotheology paradigm and supports the establishment of a civilized and sustainable society. This study emphasizes that tackling global warming necessitates not only technical and scientific solutions, but also spiritual and ethical approaches grounded in revealed values. The maqāṣidī interpretation offers an integrative and transformative analytical framework for understanding fasād as a moral failure of humanity in fulfilling the mandate of khilāfah on earth. Environmental preservation requires the integration of religious and moral dimensions as the ethical foundation for collective action within the Muslim community, with the maqāṣidī approach serving as a normative basis for Islamic ecological policies aimed at fostering sustainable ecological awareness and facilitating constructive dialogues between theology, environmental science, and public policy in comprehensively addressing the climate crisis.
Abstrak: Artikel ini bertujuan untuk merekonstruksi makna fasād (kerusakan lingkungan) dalam Al-Qur’an melalui pendekatan tafsir maqāṣidī sebagai respons teologis terhadap isu pemanasan global. Fenomena kerusakan lingkungan kontemporer seperti pencemaran, deforestasi, dan perubahan iklim ekstrem dipahami sebagai manifestasi nyata dari konsep fasād sebagaimana disebutkan dalam Q.S. Ar-Rūm: 41. Penelitian ini merupakan jenis penelitian kualitatif yang memanfaatkan sumber-sumber pustaka serta menggunakan pendekatan tafsir maqashidi sebagai pisau analisisnya. Hasil penelitian menunjukkan bahwa pemahaman terhadap ayat-ayat ekologis telah mengalami perkembangan dari tafsir klasik tekstual, menuju pendekatan ilmiah dan logis di era pertengahan, hingga tafsir kontekstual yang responsif terhadap tantangan lingkungan kontemporer. Rekonstruksi makna fasād ini memiliki relevansi penting dalam membangun paradigma ekoteologi Islam serta mendukung pembangunan peradaban yang berkeadaban dan berkelanjutan. Studi ini menegaskan bahwa penanggulangan pemanasan global tidak hanya membutuhkan solusi teknis dan ilmiah, tetapi juga pendekatan spiritual dan etis berbasis nilai-nilai wahyu. Tafsir maqāṣidī memberikan kerangka analisis integratif dan transformatif dalam memahami fasād sebagai kegagalan moral manusia dalam menjalankan mandat khilāfah di bumi. Pelestarian lingkungan menuntut integrasi dimensi religius dan moral sebagai fondasi etis dalam tindakan kolektif masyarakat Muslim, dengan pendekatan tafsir maqāṣidī yang berperan sebagai dasar normatif bagi kebijakan ekologis berwawasan Islam, guna membangun kesadaran ekologis yang berkelanjutan serta membuka ruang dialog konstruktif antara teologi, ilmu lingkungan, dan kebijakan publik dalam merespons krisis iklim secara komprehensif.
Large Language Models (LLMs) have shown impressive moral reasoning abilities. Yet they often diverge when confronted with complex, multi-factor moral dilemmas. To address these discrepancies, we propose a framework that synthesizes multiple LLMs' moral judgments into a collectively formulated moral judgment, realigning models that deviate significantly from this consensus. Our aggregation mechanism fuses continuous moral acceptability scores (beyond binary labels) into a collective probability, weighting contributions by model reliability. For misaligned models, a targeted embedding-optimization procedure fine-tunes token embeddings for moral philosophical theories, minimizing JS divergence to the consensus while preserving semantic integrity. Experiments on a large-scale social moral dilemma dataset show our approach builds robust consensus and improves individual model fidelity. These findings highlight the value of data-driven moral alignment across multiple models and its potential for safer, more consistent AI systems.
Geoffroy Morlat, Marceau Nahon, Augustin Chartouny
et al.
Moral actions are judged not only by their outcomes but by the context in which they occur. We present COMETH (Contextual Organization of Moral Evaluation from Textual Human inputs), a framework that integrates a probabilistic context learner with LLM-based semantic abstraction and human moral evaluations to model how context shapes the acceptability of ambiguous actions. We curate an empirically grounded dataset of 300 scenarios across six core actions (violating Do not kill, Do not deceive, and Do not break the law) and collect ternary judgments (Blame/Neutral/Support) from N=101 participants. A preprocessing pipeline standardizes actions via an LLM filter and MiniLM embeddings with K-means, producing robust, reproducible core-action clusters. COMETH then learns action-specific moral contexts by clustering scenarios online from human judgment distributions using principled divergence criteria. To generalize and explain predictions, a Generalization module extracts concise, non-evaluative binary contextual features and learns feature weights in a transparent likelihood-based model. Empirically, COMETH roughly doubles alignment with majority human judgments relative to end-to-end LLM prompting (approx. 60% vs. approx. 30% on average), while revealing which contextual features drive its predictions. The contributions are: (i) an empirically grounded moral-context dataset, (ii) a reproducible pipeline combining human judgments with model-based context learning and LLM semantics, and (iii) an interpretable alternative to end-to-end LLMs for context-sensitive moral prediction and explanation.
Rejections of positive offers in the Ultimatum Game have been attributed to different motivations. We show that a model combining social preferences and moral concerns provides a unifying explanation for these rejections while accounting for additional evidence. Under the preferences considered, a positive degree of spite is a necessary and sufficient condition for rejecting positive offers. This indicates that social preferences, rather than moral concerns, drive rejection behavior. This does not imply that moral concerns do not matter. We show that rejection thresholds increase with individuals' moral concerns, suggesting that morality acts as an amplifier of social preferences. Using data from van Leeuwen and Alger (2024), we estimate individuals' social preferences and moral concerns using a finite mixture approach. Consistent with previous evidence, we identify two types of individuals who reject positive offers in the Ultimatum Game, but that differ in their Dictator Game behavior.
The proliferation of Artificial Intelligence (AI) systems exhibiting complex and seemingly agentive behaviours necessitates a critical philosophical examination of their agency, autonomy, and moral status. In this paper we undertake a systematic analysis of the differences between basic, autonomous, and moral agency in artificial systems. We argue that while current AI systems are highly sophisticated, they lack genuine agency and autonomy because: they operate within rigid boundaries of pre-programmed objectives rather than exhibiting true goal-directed behaviour within their environment; they cannot authentically shape their engagement with the world; and they lack the critical self-reflection and autonomy competencies required for full autonomy. Nonetheless, we do not rule out the possibility of future systems that could achieve a limited form of artificial moral agency without consciousness through hybrid approaches to ethical decision-making. This leads us to suggest, by appealing to the necessity of consciousness for moral patiency, that such non-conscious AMAs might represent a case that challenges traditional assumptions about the necessary connection between moral agency and moral patiency.
Large language models (LLMs) increasingly operate in social contexts, motivating analysis of how they express and shift moral judgments. In this work, we investigate the moral response of LLMs to persona role-play, prompting a LLM to assume a specific character. Using the Moral Foundations Questionnaire (MFQ), we introduce a benchmark that quantifies two properties: moral susceptibility and moral robustness, defined from the variability of MFQ scores across and within personas, respectively. We find that, for moral robustness, model family accounts for most of the variance, while model size shows no systematic effect. The Claude family is, by a significant margin, the most robust, followed by Gemini and GPT-4 models, with other families exhibiting lower robustness. In contrast, moral susceptibility exhibits a mild family effect but a clear within-family size effect, with larger variants being more susceptible. Moreover, robustness and susceptibility are positively correlated, an association that is more pronounced at the family level. Additionally, we present moral foundation profiles for models without persona role-play and for personas averaged across models. Together, these analyses provide a systematic view of how persona conditioning shapes moral behavior in LLMs.
Jea Kwon, Luiz Felipe Vecchietti, Sungwon Park
et al.
Humans display significant uncertainty when confronted with moral dilemmas, yet the extent of such uncertainty in machines and AI agents remains underexplored. Recent studies have confirmed the overly confident tendencies of machine-generated responses, particularly in large language models (LLMs). As these systems are increasingly embedded in ethical decision-making scenarios, it is important to understand their moral reasoning and the inherent uncertainties in building reliable AI systems. This work examines how uncertainty influences moral decisions in the classical trolley problem, analyzing responses from 32 open-source models and 9 distinct moral dimensions. We first find that variance in model confidence is greater across models than within moral dimensions, suggesting that moral uncertainty is predominantly shaped by model architecture and training method. To quantify uncertainty, we measure binary entropy as a linear combination of total entropy, conditional entropy, and mutual information. To examine its effects, we introduce stochasticity into models via "dropout" at inference time. Our findings show that our mechanism increases total entropy, mainly through a rise in mutual information, while conditional entropy remains largely unchanged. Moreover, this mechanism significantly improves human-LLM moral alignment, with correlations in mutual information and alignment score shifts. Our results highlight the potential to better align model-generated decisions and human preferences by deliberately modulating uncertainty and reducing LLMs' confidence in morally complex scenarios.
Large language models (LLMs) are increasingly deployed in domains requiring moral understanding, yet their reasoning often remains shallow, and misaligned with human reasoning. Unlike humans, whose moral reasoning integrates contextual trade-offs, value systems, and ethical theories, LLMs often rely on surface patterns, leading to biased decisions in morally and ethically complex scenarios. To address this gap, we present a value-grounded framework for evaluating and distilling structured moral reasoning in LLMs. We benchmark 12 open-source models across four moral datasets using a taxonomy of prompts grounded in value systems, ethical theories, and cognitive reasoning strategies. Our evaluation is guided by four questions: (1) Does reasoning improve LLM decision-making over direct prompting? (2) Which types of value/ethical frameworks most effectively guide LLM reasoning? (3) Which cognitive reasoning strategies lead to better moral performance? (4) Can small-sized LLMs acquire moral competence through distillation? We find that prompting with explicit moral structure consistently improves accuracy and coherence, with first-principles reasoning and Schwartz's + care-ethics scaffolds yielding the strongest gains. Furthermore, our supervised distillation approach transfers moral competence from large to small models without additional inference cost. Together, our results offer a scalable path toward interpretable and value-grounded models.
Linda Zagzebski’s exemplarist moral theory (EMT) has much to commend it, but without appeal to a single, paradigmatic exemplar, it remains vulnerable to epistemic issues, such as: How do we reliably distinguish between who is admirable and who is not? In this paper, I argue that a Christocentric version of her theory is capable of addressing this problem. The paper’s aims are: (1) to demonstrate how teleologically rooting EMT in Christ helps address its epistemic issues, as related specifically to the relationship between individuality and universality, and (2) to present in vivid detail the Christian moral exemplar, using Zagzebski’s framework, as a means to both further the first aim and develop its implications. This is achieved through a kind of case study of three (corresponding, I argue) biblical accounts of ascent: Abraham up Moriah, Moses up Sinai, and Christ up Tabor, with special attention given to Kierkegaard’s interpretation of Abraham and Dionysius’s interpretation of Moses. Following this case study, I elucidate the virtue, motive, end, and act specific to the Christian moral exemplar.
La cuestión de la relevancia actual del cristianismo aparece en distintos debates sobre la comprensión de la Iglesia católica y sobre su doctrina teológica y moral. Esta discusión se ha agudizado, sobre todo en la teología alemana, con una controversia sobre la relación de la libertad con la verdad. En este artículo se presenta el contenido fundamental de esa controversia y se reflexiona sobre algunos aspectos de la misma, como la relación entre libertad y amor, la idea de auto-determinación y la moralidad. Finalmente, se apuntan algunas reflexiones que ayuden a pensar un concepto de libertad, insistiendo en su caráter global y en la capacidad creativa propia del amor.
Abstract: The question of the current relevance of Christianity appears in different debates on the understanding of the Catholic Church and on its theological and moral doctrine. This discussion has sharpened, specially in German theology, with controversy about the relation of freedom to truth. This article presents the fundamental content of this controversy and reflects on some aspects of it, such as the relationship between freedom and love the idea of self-determination and the morality. Finally, some reflections are pointed out that help to think about a concept of freedom, insisting on its global character and on the creative capacity of love.
Preference elicitation frameworks feature heavily in the research on participatory ethical AI tools and provide a viable mechanism to enquire and incorporate the moral values of various stakeholders. As part of the elicitation process, surveys about moral preferences, opinions, and judgments are typically administered only once to each participant. This methodological practice is reasonable if participants' responses are stable over time such that, all other relevant factors being held constant, their responses today will be the same as their responses to the same questions at a later time. However, we do not know how often that is the case. It is possible that participants' true moral preferences change, are subject to temporary moods or whims, or are influenced by environmental factors we don't track. If participants' moral responses are unstable in such ways, it would raise important methodological and theoretical issues for how participants' true moral preferences, opinions, and judgments can be ascertained. We address this possibility here by asking the same survey participants the same moral questions about which patient should receive a kidney when only one is available ten times in ten different sessions over two weeks, varying only presentation order across sessions. We measured how often participants gave different responses to simple (Study One) and more complicated (Study Two) repeated scenarios. On average, the fraction of times participants changed their responses to controversial scenarios was around 10-18% across studies, and this instability is observed to have positive associations with response time and decision-making difficulty. We discuss the implications of these results for the efficacy of moral preference elicitation, highlighting the role of response instability in causing value misalignment between stakeholders and AI tools trained on their moral judgments.
Moral values play a fundamental role in how we evaluate information, make decisions, and form judgements around important social issues. The possibility to extract morality rapidly from lyrics enables a deeper understanding of our music-listening behaviours. Building on the Moral Foundations Theory (MFT), we tasked a set of transformer-based language models (BERT) fine-tuned on 2,721 synthetic lyrics generated by a large language model (GPT-4) to detect moral values in 200 real music lyrics annotated by two experts.We evaluate their predictive capabilities against a series of baselines including out-of-domain (BERT fine-tuned on MFT-annotated social media texts) and zero-shot (GPT-4) classification. The proposed models yielded the best accuracy across experiments, with an average F1 weighted score of 0.8. This performance is, on average, 5% higher than out-of-domain and zero-shot models. When examining precision in binary classification, the proposed models perform on average 12% higher than the baselines.Our approach contributes to annotation-free and effective lyrics morality learning, and provides useful insights into the knowledge distillation of LLMs regarding moral expression in music, and the potential impact of these technologies on the creative industries and musical culture.
Vijay Keswani, Vincent Conitzer, Hoda Heidari
et al.
Computational preference elicitation methods are tools used to learn people's preferences quantitatively in a given context. Recent works on preference elicitation advocate for active learning as an efficient method to iteratively construct queries (framed as comparisons between context-specific cases) that are likely to be most informative about an agent's underlying preferences. In this work, we argue that the use of active learning for moral preference elicitation relies on certain assumptions about the underlying moral preferences, which can be violated in practice. Specifically, we highlight the following common assumptions (a) preferences are stable over time and not sensitive to the sequence of presented queries, (b) the appropriate hypothesis class is chosen to model moral preferences, and (c) noise in the agent's responses is limited. While these assumptions can be appropriate for preference elicitation in certain domains, prior research on moral psychology suggests they may not be valid for moral judgments. Through a synthetic simulation of preferences that violate the above assumptions, we observe that active learning can have similar or worse performance than a basic random query selection method in certain settings. Yet, simulation results also demonstrate that active learning can still be viable if the degree of instability or noise is relatively small and when the agent's preferences can be approximately represented with the hypothesis class used for learning. Our study highlights the nuances associated with effective moral preference elicitation in practice and advocates for the cautious use of active learning as a methodology to learn moral preferences.
Hong Zhang, Quoc-Nam Nguyen, Prasanta Bhattacharya
et al.
This study enhances stance detection on social media by incorporating deeper psychological attributes, specifically individuals' moral foundations. These theoretically-derived dimensions aim to provide a comprehensive profile of an individual's moral concerns which, in recent work, has been linked to behaviour in a range of domains, including society, politics, health, and the environment. In this paper, we investigate how moral foundation dimensions can contribute to predicting an individual's stance on a given target. Specifically we incorporate moral foundation features extracted from text, along with message semantic features, to classify stances at both message- and user-levels using both traditional machine learning models and large language models. Our preliminary results suggest that encoding moral foundations can enhance the performance of stance detection tasks and help illuminate the associations between specific moral foundations and online stances on target topics. The results highlight the importance of considering deeper psychological attributes in stance analysis and underscores the role of moral foundations in guiding online social behavior.