We study sentence-level detection of the 19 human values in the refined Schwartz continuum in about 74k English sentences from news and political manifestos (ValueEval'24 corpus). Each sentence is annotated with value presence, yielding a binary moral-presence label and a 19-way multi-label task under severe class imbalance. First, we show that moral presence is learnable from single sentences: a DeBERTa-base classifier attains positive-class F1 = 0.74 with calibrated thresholds. Second, we compare direct multi-label value detectors with presence-gated hierarchies in a setting where only a single consumer-grade GPU with 8 GB of VRAM is available, and we explicitly choose all training and inference configurations to fit within this budget. Presence gating does not improve over direct prediction, indicating that gate recall becomes a bottleneck. Third, we investigate lightweight auxiliary signals - short-range context, LIWC-22, and moral lexica - and small ensembles. Our best supervised configuration, a soft-voting ensemble of DeBERTa-based models enriched with such signals, reaches macro-F1 = 0.332 on the 19 values, improving over the best previous English-only baseline on this corpus, namely the best official ValueEval'24 English run (macro-F1 = 0.28 on the same 19-value test set). Methodologically, our study provides, to our knowledge, the first systematic comparison of direct versus presence-gated architectures, lightweight feature-augmented encoders, and medium-sized instruction-tuned Large Language Models (LLMs) for refined Schwartz values at sentence level. We additionally benchmark 7-9B instruction-tuned LLMs (Gemma 2 9B, Llama 3.1 8B, Mistral 8B, Qwen 2.5 7B) in zero-/few-shot and QLoRA setups, and find that they lag behind the supervised ensemble under the same compute budget. Overall, our results provide empirical guidance for building compute-efficient, value-aware NLP models.
Do We Need Role Models? How do Role Models Shape Collective Morality? To explore the questions, we build a multi-agent simulation powered by a Large Language Model, where agents with diverse intrinsic drives, ranging from cooperative to competitive, interact and adapt through a four-stage cognitive loop (plan-act-observe-reflect). We design four experimental games (Alignment, Collapse, Conflict, and Construction) and conduct motivational ablation studies to identify the key drivers of imitation. The results indicate that identity-driven conformity can powerfully override initial dispositions. Agents consistently adapt their values to align with a perceived successful exemplar, leading to rapid value convergence.
Taufiqur Rohman, Agus Imam Kharomen, Hayim Muhammad
This study analyzes the narrative of Ashābul Fīl in Surah al-Fīl through Muhammad Ahmad Khalafullah's al-fann al-qashāṣī framework. While classical exegesis largely emphasizes the historical dimension of the event, limited research systematically applies Khalafullah's literary-narrative approach to short Makkiyyah surahs. This study addresses that gap by examining the aesthetic structure, symbolism, and theological meaning of QS. al-Fīl beyond historiographical concerns. Using a qualitative-descriptive design with a literary-hermeneutic approach, the research is based on textual analysis of the Qur'an, classical and modern tafsīr, and Khalafullah's theoretical works. Narrative structural analysis and thematic interpretation are employed to identify symbolic and theological patterns within the text. The findings show that QS. al-Fīl presents a dramatic structure consisting of threat, divine intervention, and the destruction of tyranny. The symbols of ṭayran abābīl and ḥijārah min sijjīl function as theological metaphors representing divine sovereignty and moral accountability. This study demonstrates that Khalafullah's approach enriches contemporary Qur'anic interpretation by integrating aesthetics, symbolism, and theology. Globally, the research highlights the Qur'an's ethical relevance in addressing injustice, authoritarianism, and moral crisis in modern society.
Nowadays, social media is pivotal in shaping public discourse, especially on polarizing issues like vaccination, where diverse moral perspectives influence individual opinions. In NLP, data scarcity and complexity of psycholinguistic tasks, such as identifying morality frames, make relying solely on human annotators costly, time-consuming, and prone to inconsistency due to cognitive load. To address these issues, we leverage large language models (LLMs), which are adept at adapting new tasks through few-shot learning, utilizing a handful of in-context examples coupled with explanations that connect examples to task principles. Our research explores LLMs' potential to assist human annotators in identifying morality frames within vaccination debates on social media. We employ a two-step process: generating concepts and explanations with LLMs, followed by human evaluation using a "think-aloud" tool. Our study shows that integrating LLMs into the annotation process enhances accuracy, reduces task difficulty, lowers cognitive load, suggesting a promising avenue for human-AI collaboration in complex psycholinguistic tasks.
Shahriyar Zaman Ridoy, Azmine Toushik Wasi, Koushik Ahamed Tonmoy
As multilingual Large Language Models (LLMs) gain traction across South Asia, their alignment with local ethical norms, particularly for Bengali, which is spoken by over 285 million people and ranked 6th globally, remains underexplored. Existing ethics benchmarks are largely English-centric and shaped by Western frameworks, overlooking cultural nuances critical for real-world deployment. To address this, we introduce BengaliMoralBench, the first large-scale ethics benchmark for the Bengali language and socio-cultural contexts. It covers five moral domains, Daily Activities, Habits, Parenting, Family Relationships, and Religious Activities, subdivided into 50 culturally relevant subtopics. Each scenario is annotated via native-speaker consensus using three ethical lenses: Virtue, Commonsense, and Justice ethics. We conduct systematic zero-shot evaluation of prominent multilingual LLMs, including Llama, Gemma, Qwen, and DeepSeek, using a unified prompting protocol and standard metrics. Performance varies widely (50-91% accuracy), with qualitative analysis revealing consistent weaknesses in cultural grounding, commonsense reasoning, and moral fairness. BengaliMoralBench provides a foundation for responsible localization, enabling culturally aligned evaluation and supporting the deployment of ethically robust AI in diverse, low-resource multilingual settings such as Bangladesh.
Large language models (LLMs) have been shown to internalize human-like biases during finetuning, yet the mechanisms by which these biases manifest remain unclear. In this work, we investigated whether the well-known Knobe effect, a moral bias in intentionality judgements, emerges in finetuned LLMs and whether it can be traced back to specific components of the model. We conducted a Layer-Patching analysis across 3 open-weights LLMs and demonstrated that the bias is not only learned during finetuning but also localized in a specific set of layers. Surprisingly, we found that patching activations from the corresponding pretrained model into just a few critical layers is sufficient to eliminate the effect. Our findings offer new evidence that social biases in LLMs can be interpreted, localized, and mitigated through targeted interventions, without the need for model retraining.
As AI hype continues to grow, organizations face pressure to broadcast or downplay purported AI initiatives - even when contrary to truth. This paper introduces AI-washing as overstating (deceptive boasting) or understating (deceptive denial) a company's real AI usage. A 2x2 experiment (N = 401) examines how these false claims affect consumer attitudes and purchase intentions. Results reveal a pronounced asymmetry: deceptive denial evokes more negative moral judgments than honest negation, while deceptive boasting has no effects. We show that perceived betrayal mediates these outcomes. By clarifying how AI-washing erodes trust, the study highlights clear ethical implications for policymakers, marketers, and researchers striving for transparency.
The article presents new entropic continuity bounds for conditional expectations and conditional covariance matrices. These bounds are expressed in terms of the relative entropy between different coupling distributions. Our approach combines Wasserstein coupling with quadratic transportation cost inequalities. We illustrate the impact of these results in the context of entropic optimal transport problems. The entropic continuity theorem presented in the article allows to estimate the conditional expectations and the conditional covariances of Schr\" odinger and Sinkhorn transitions in terms of the relative entropy between the corresponding bridges. These entropic continuity bounds turns out to be a very useful tool for obtaining remarkably simple proofs of the exponential decays of the gradient and the Hessian of Schrödinger and Sinkhorn bridge potentials.
Oscar Araque, Lorenzo Gatti, Sergio Consoli
et al.
The moral value of liberty is a central concept in our inference system when it comes to taking a stance towards controversial social issues such as vaccine hesitancy, climate change, or the right to abortion. Here, we propose a novel Liberty lexicon evaluated on more than 3,000 manually annotated data both in in- and out-of-domain scenarios. As a result of this evaluation, we produce a combined lexicon that constitutes the main outcome of this work. This final lexicon incorporates information from an ensemble of lexicons that have been generated using word embedding similarity (WE) and compositional semantics (CS). Our key contributions include enriching the liberty annotations, developing a robust liberty lexicon for broader application, and revealing the complexity of expressions related to liberty across different platforms. Through the evaluation, we show that the difficulty of the task calls for designing approaches that combine knowledge, in an effort of improving the representations of learning systems.
Hinduizm’de ibadet genellikle bireysel bir şekilde gerçekleştirilmekle birlikte topluca uygulanan biçimleri de bulunmaktadır. Yıllık festivaller toplu ibadetlerin en belirgin örnekleridir. Dini ve kültürel yaşamın önemli bir parçasını oluşturan bu festivallerin her biri farklı tanrılarla yahut tanrıçalarla ve mitolojik olaylarla ilişkilidir. Hindu festivalleri ritüelleri, duaları, dansları, festivalle ilişkili heykellerin yapılmasını, kutlamanın konusuyla ilgili olayların sergilendiği çadırların hazırlanmasını ve birlikte yenilen yemekleri içerir. Birlik ve beraberlik içinde kutlanan festivaller bir yandan tapınılan tanrıya, tanrıçaya bağlılığı kuvvetlendirirken diğer yandan toplumsal birliği pekiştirir. Hinduizm’deki festivallerin kutlama biçimleri bölgeden bölgeye değişiklik gösterdiği gibi dini geleneklere göre de farklılıklar ihtiva eder. Çalışmamızın konusu olan Durgā Pūcā ve Kālī Pūcā festivalleri tüm Hindular tarafından kutlanan önemli bir festivaldir. Tanrıçaya adanan bu festivallerden Durgā Pūcā tüm Hindistan’da Kālī Pūcā ise Batı Bengal, Assam, Odişa/Orissa ve Tripura gibi doğu Hindistan bölgelerinde kutlanır. Makale Durgā Pūcā ve Kālī Pūcā festivallerini ve bu festivallerde yer alan tantracı unsurları incelemeyi amaçlar. İnceleme konusu olarak bu iki festivalin seçilmesinin sebebi genel olarak Hinduizm’deki temel kutlama şekillerini devam ettirmekle beraber tanrıçaya tapımın ve hayvan kurbanının festivalin merkezinde yer almasıdır. Makalede Durgā Pūcā ve Kālī Pūcā festivallerinden söz edildikten sonra Tantracılığın ne zaman ve nerede ortaya çıktığından ve tantracı unsurlardan bahsedilecektir. Durgā Pūcā ve Kālī Pūcā festivallerinde tantracı unsurların nasıl yer aldığı ve bu unsurların dini açıdan ne anlama geldiği üzerinde durulacaktır. Bu bağlamda tanrıçaya tapım ve hayvan kurbanının bu festivallerin merkezinde yer alması tantracı düşüncenin etkisini anlamak açısından önemlidir. Zira Tantracılık yüce varlık anlayışında şaktiye önem verir ve onun tezahürleri olarak görülen tanrıçayı merkeze koyar, geleneksel Hindu düşüncesinin kabul etmediği ibadetleri kabul eder ve uygular. Zaman içerisinde bu ibadetler belirli gruplar tarafından devam ettirilmekle birlikte bazı ibadetlerin ve festivallerin kutlanış biçimlerinde kendini gösterir. Çalışmada dinler tarihinin betimleyici yöntemi kullanılarak festivaller, kavramlar ve fenomenler tanımlanmaya çalışılmıştır. Din fenomenolojisi metodundan faydalanarak Durgā Pūcā ve Kālī Pūcā festivallerinde gerçekleştirilen ritüellerin incelemesi yapılmıştır. Kavramların Türkçe’ye doğru ve eksiksiz bir şekilde aktarılabilmesi için Türkçe’de karşılığı olan harflerin yanında Hint hükümeti tarafından kabul edilen ISO 15019 çeviri yazım tablosu kullanılmıştır.
With Artificial Intelligence systems increasingly applied in consequential domains, researchers have begun to ask how these systems ought to act in ethically charged situations where even humans lack consensus. In the Moral Machine project, researchers crowdsourced answers to "Trolley Problems" concerning autonomous vehicles. Subsequently, Noothigattu et al. (2018) proposed inferring linear functions that approximate each individual's preferences and aggregating these linear models by averaging parameters across the population. In this paper, we examine this averaging mechanism, focusing on fairness concerns in the presence of strategic effects. We investigate a simple setting where the population consists of two groups, with the minority constituting an α < 0.5 share of the population. To simplify the analysis, we consider the extreme case in which within-group preferences are homogeneous. Focusing on the fraction of contested cases where the minority group prevails, we make the following observations: (a) even when all parties report their preferences truthfully, the fraction of disputes where the minority prevails is less than proportionate in α; (b) the degree of sub-proportionality grows more severe as the level of disagreement between the groups increases; (c) when parties report preferences strategically, pure strategy equilibria do not always exist; and (d) whenever a pure strategy equilibrium exists, the majority group prevails 100% of the time. These findings raise concerns about stability and fairness of preference vector averaging as a mechanism for aggregating diverging voices. Finally, we discuss alternatives, including randomized dictatorship and median-based mechanisms.
This article reflects on the apologies by Pope Francis for Indian Residential Schools in Canada in terms of the stories they tell about past wrongdoing and about the role of the apologies in repair and healing. Public commentators on apologies such as myself are part of this narrative process. These papal apologies frame the integrity of the representative making the apology, the scope of wrongdoing, and thus which kinds of actions the church believes will advance reconciliation. While an analysis of the words spoken reveal significant claims, including the Catholic Church’s problematic denial of institutional responsibility for the schools and the policies they advanced, the meaning of an apology is not determined only by an analysis of the words. I argue that the transformative promise of a praxis which includes an apology is linked to the apologizer’s renunciation of control over the process.
Text-to-image generation methods produce high-resolution and high-quality images, but these methods should not produce immoral images that may contain inappropriate content from the perspective of commonsense morality. In this paper, we aim to automatically judge the immorality of synthesized images and manipulate these images into morally acceptable alternatives. To this end, we build a model that has three main primitives: (1) recognition of the visual commonsense immorality in a given image, (2) localization or highlighting of immoral visual (and textual) attributes that contribute to the immorality of the image, and (3) manipulation of an immoral image to create a morally-qualifying alternative. We conduct experiments and human studies using the state-of-the-art Stable Diffusion text-to-image generation model, demonstrating the effectiveness of our ethical image manipulation approach.
Al-Ījī presents the final stage of Ashʿarism, and his arguments reflect the traditional and philosophical approaches in the school. This paper presents the main arguments that al-Ījī deployed to refute the Muʿtazilites’ ethical realism. Its aim is to present the exact form of al-Ījī’s arguments, explain them, discuss the objections, and then evaluate their strengths. The paper’s aim is to explain the Muʿtazilites’ arguments; nevertheless, it gives some clarifications when it is needed to understand al-Ījī’s arguments in a better way. In the beginning, the paper draws a distinct line between the Ashʿarites’ and the Muʿtazilite ethical understandings. It defines the focus of the controversy and prepares the groundwork for theological arguments. Al-Ījī’s arguments are divided into three categories. In the first one, we present al-Ījī’s general argument, which is a polemic argument whereby al-Ījī tries to negate the freedom of human choice in order to cast doubt on the Muʿtazilite ethical foundations. We discuss its critique, and then reveal al-Ījī’s real position on human power and freedom of choice. The second category contains three arguments against the intrinsic ethical value: two of the arguments were adopted by al-Ījī and the third was attributed to other Ashʿarites in a general way. The first two arguments deal with the intrinsic ethical values of lying and truth-telling, while the third one is based on the Ashʿarite famous assertion: ‘an accident cannot subsist on another accident.’ The final category is dedicated to discussing al-Ījī’s argument against the Muʿtazilite theory of ethical aspects. A sufficient account of the theory and its partisans is provided before discussing al-Ījī’s argument. Moreover, a brief introduction of Al-Ījī’s and the Muʿtazilites’ conception of divine ethics is discussed in the folds of the argument. Some divine qualities, such as justice and wisdom, are defined from the Ashʿarites’ and the Muʿtazilites’ perspectives. As a result, the paper gives a clear account of al-Ījī’s arguments against the Muʿtazilites’ ethical realism; it presents and evaluates the objections and defines the strengths and the defects in the arguments. Finally, it proposes a better way to understand the Ashʿarites’ ethical arguments in their right context.
The idea of the common good is missing in politics today. Fighting for political ideology and self-interest has replaced finding solutions to problems or practising the ethics of public interest. We urgently need to create a new social contract with proper implementation of the values set out in the national Constitution. This study was undertaken from a reformed ethical perspective, with special emphasis on the ethics of the common good. Methodologically, in researching this article, the author was guided by two related questions: how do ideas, beliefs and norms form? What happens in society to let these norms shape our actions? Conducting research guided by these questions has helped the author to understand that for many communities, stability is maintained by rules, norms, beliefs, convictions and worldview as located in tradition and culture. It is institutions such as those outlined here that still guide attitude and behaviour in the majority of cases. While being sensitive to these institutions and the role they play, through policy and legislation, under constitutional supremacy, the Constitution has become the primary guide and source for community stability.
Interdisciplinary and/or interdisciplinary implications: This article brings the disciplines of theology, politics and governance together in defining the ethics of the common good in contemporary South African politics. It proposes that the moral prerequisite for solving the deepest problems our country now face is a commitment to the ethics of the common good. By definition, this will require the engagement and collaboration of all the ‘stakeholders’ – government, businesses, civil society groups, faith groups and especially young people.
This paper explores the impact of the Great War (1914–1918) on Christianity in Scotland. This includes the attitudes of various denominations to the war, the rise and fall of religious attendance, and the impact of the war on Sabbatarianism, Sunday schools, church attendance, theology, and social and moral issues in Scottish society.
Markus Peschl, Arkady Zgonnikov, Frans A. Oliehoek
et al.
Inferring reward functions from demonstrations and pairwise preferences are auspicious approaches for aligning Reinforcement Learning (RL) agents with human intentions. However, state-of-the art methods typically focus on learning a single reward model, thus rendering it difficult to trade off different reward functions from multiple experts. We propose Multi-Objective Reinforced Active Learning (MORAL), a novel method for combining diverse demonstrations of social norms into a Pareto-optimal policy. Through maintaining a distribution over scalarization weights, our approach is able to interactively tune a deep RL agent towards a variety of preferences, while eliminating the need for computing multiple policies. We empirically demonstrate the effectiveness of MORAL in two scenarios, which model a delivery and an emergency task that require an agent to act in the presence of normative conflicts. Overall, we consider our research a step towards multi-objective RL with learned rewards, bridging the gap between current reward learning and machine ethics literature.
Hz. Peygamber’in vefatı sonrası Hz. Osman’ın (öl. 35/656) şehit edilmesi ve Hz. Ali’nin (öl. 40/661) halifeliği döneminde vuku bulan Cemel ve Sıffîn savaşları özellikle Müslümanları karşı karşıya getirmiştir. Yaşanan bu hâdiselerden sonra başlayan tartışmalar, özelde bir Müslümanı öldürenin durumu genelde ise kader ekseninde sürdürülmüştür. İlk dönemde ortaya çıkan bireysel düşünceler zamanla fırka/mezheplerin görüşlerine dönüşmüş, bu fırka/mezhepler diğer konularda olduğu gibi bu konuda da birbirlerini tenkit etmişler, konuyla ilgili reddiyeler kaleme almışlardır. Bu köklü ve çetrefilli konu, günümüze kadar gelen süreçte, Osmanlı’da da ele alınmış, Arapça ve Türkçe olarak ya eserlerin içinde bir bölüm olarak yer almış ya da müstakil risâleler halinde ele alınmıştır. Bu çalışmanın konusu olan Boyabatlı Halil Şükrü Efendi’nin (öl.?) kazâ ve kader hakkındaki risalesi de işte bunlardan biridir. Müellif müstakil olarak kaleme aldığı risâlesinde, kendi düşüncelerini örneklerle açıklarken bazı fırkaların kazâ ve kader hakkındaki görüşlerini de çürütmeye çalışmıştır. Bu çalışmada Hafız Halil Şükrü Efendi’nin kazâ ve kader hakkındaki risalesi temel alınmıştır. Müellifin hayatı ve ilmî şahsiyeti hakkında ulaşılan arşiv belgelerinden özet bilgiler verilmiş, daha sonra risale tahlil edilmeye çalışılmıştır. Risalede müellif tarafından herhangi bir alt başlık kullanılmamıştır. Ancak risalenin tahlilinde tarafımızdan konunun daha rahat anlaşılmasını sağlamak amacıyla ilgili yerlerde alt başlıklar konulmuştur. Kazâ ve kader kavramları, fiillerin Allah’a ve insana nispet edilmesi, kaderin önceden belirli olup olmadığı, Cebriyye ve Muʻtezile fırkalarına verdiği cevaplar bu başlıklardan birkaç tanesidir. Halil Şükrü Efendi insan fiillerini kesb ve halk, kazâ ve makzî olarak iki ayrı grupta incelemiş, fiillerin failiyet mahalliyet açısından analizini de yapmaya çalışmıştır. Kader üst başlığında ele alınan hidâyet ve dalâlet ile adâlet ve zulüm konularında da görüşlerini belirten müellif teklif noktasında ilgi çekici değerlendirmelerde bulunmuştur. Halil Şükrü Efendi insan iradesi ile teklifin ayrı şeyler olduğunu, biri olduğu için diğerinin var olduğunu değil ikisinin de bağımsız olarak var olduğunu söylemiştir. Yani teklif olduğu için iman veya küfrün olmayacağını dile getiren Halil Şükrü Efendi’ye göre teklifin bu noktada insan iradesi veya seçimleri üzerinde bir bağlayıcılığı yoktur. Buna göre Halil Şükrü Efendi’nin, insan iradesiyle ilgili olarak Ebû Mansûr el-Mâtürîdî’nin (öl. 333/944) görüşlerine yakın bir görüşte olduğu söylenebilir.