mlsanityData to 7 Oct 2026

Language Model Alignment in Multilingual Trolley Problems

arXiv 2407.02273 ↗First posted 2 Jul 2024cs.CLcited by 46

On the map: LLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment

Abstract (arXiv)

We evaluate the moral alignment of LLMs with human preferences in multilingual trolley problems. Building on the Moral Machine experiment, which captures over 40 million human judgments across 200+ countries, we develop a cross-lingual corpus of moral dilemma vignettes in over 100 languages called MultiTP. This dataset enables the assessment of LLMs' decision-making processes in diverse linguistic contexts. Our analysis explores the alignment of 19 different LLMs with human judgments, capturing preferences across six moral dimensions: species, gender, fitness, status, age, and the number of lives involved. By correlating these preferences with the demographic distribution of language speakers and examining the consistency of LLM responses to various prompt paraphrasings, our findings provide insights into cross-lingual and ethical biases of LLMs and their intersection. We discover significant variance in alignment across languages, challenging the assumption of uniform moral reasoning in AI systems and highlighting the importance of incorporating diverse perspectives in AI ethics. The results underscore the need for further research on the integration of multilingual dimensions in responsible AI research to ensure fair and equitable AI interactions worldwide. Our code and data are at https://github.com/causalNLP/moralmachine

Key facts. Cited by 46 papers in the map (data through 7 Oct 2026). Placed in Human values and AI feedback alignment, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 56 of its 85 references that are in the map. First posted on arXiv 2 Jul 2024.

Builds on

56 of 85 references.

Closest ancestors · 5

Same direction, newest first

  1. 2024Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference7 Mar
  2. 2024A Roadmap to Pluralistic Alignment7 Feb
  3. 2023MoCa: Measuring Human-Language Model Alignment on Causal and Moral Judgment Tasks30 Oct
  4. 2022Constitutional AI: Harmlessness from AI Feedback15 Dec
  5. 2020Aligning AI With Shared Human Values5 Aug

Other references · 51

  1. 2020Language Models are Few-Shot Learnerscited by 32,199
  2. 2023GPT-4 Technical Reportcited by 18,132
  3. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  4. 2021On the Opportunities and Risks of Foundation Modelscited by 3,961
  5. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,383
  6. 2023Qwen Technical Reportcited by 3,273
  7. 2023Mistral 7Bcited by 3,065
  8. 2023Sparks of Artificial General Intelligence: Early experiments with GPT-4cited by 2,579
  9. 2024Mixtral of Expertscited by 1,785
  10. 2023Jailbroken: How Does LLM Safety Training Fail?cited by 1,565
  11. 2022No Language Left Behind: Scaling Human-Centered Machine Translationcited by 1,290
  12. 2024Refusal in Language Models Is Mediated by a Single Directioncited by 960
38 more
  1. 2018XNLI: Evaluating Cross-lingual Sentence Representationscited by 913
  2. 2023Whose Opinions Do Language Models Reflect?cited by 759
  3. 2024Yi: Open Foundation Models by 01.AIcited by 706
  4. 2019On the Cross-lingual Transferability of Monolingual Representationscited by 575
  5. 2021The FLORES-101 Evaluation Benchmark for Low-Resource and Multilingual Machine Translationcited by 556
  6. 2023Towards Measuring the Representation of Subjective Global Opinions in Language Modelscited by 424
  7. 2023Are aligned neural networks adversarially aligned?cited by 305
  8. 2021Unsolved Problems in ML Safetycited by 291
  9. 2023MEGA: Multilingual Evaluation of Generative AIcited by 291
  10. 2020Social Chemistry 101: Learning to Reason about Social and Moral Normscited by 253
  11. 2023Evaluating the Moral Beliefs Encoded in LLMscited by 211
  12. 2024Foundational Challenges in Assuring Alignment and Safety of Large Language Modelscited by 193
  13. 2021Large Pre-trained Language Models Contain Human-like Biases of What is Right and Wrong to Docited by 178
  14. 2023Is ChatGPT A Good Translator? Yes With GPT-4 As The Enginecited by 172
  15. 2024Towards Measuring and Modeling "Culture" in LLMs: A Surveycited by 166
  16. 2023Multilingual Machine Translation with Large Language Models: Empirical Results and Analysiscited by 156
  17. 2020MKQA: A Linguistically Diverse Benchmark for Multilingual Open Domain Question Answeringcited by 151
  18. 2024How do Large Language Models Handle Multilingualism?cited by 148
  19. 2021Can Machines Learn Morality? The Delphi Experimentcited by 140
  20. 2023Towards Making the Most of ChatGPT for Machine Translationcited by 138
  21. 2022Jury Learning: Integrating Dissenting Voices into Machine Learning Modelscited by 115
  22. 2024CultureBank: An Online Community-Driven Knowledge Base Towards Culturally Aware Language Technologiescited by 98
  23. 2023Value Kaleidoscope: Engaging AI with Pluralistic Human Values, Rights, and Dutiescited by 97
  24. 2022When to Make Exceptions: Exploring Language Models as Accounts of Human Moral Judgmentcited by 93
  25. 2024Large Language Models are Geographically Biasedcited by 91
  26. 2023Multilingual LLMs are Better Cross-lingual In-context Learners with Alignmentcited by 71
  27. 2024Unintended Impacts of LLM Alignment on Global Representationcited by 66
  28. 2024Understanding the Capabilities and Limitations of Large Language Models for Cultural Commonsensecited by 50
  29. 2023Managing extreme AI risks amid rapid progresscited by 48
  30. 2023Democratizing LLMs for Low-Resource Languages by Leveraging their English Dominant Abilities with Linguistically-Diverse Promptscited by 45
  31. 2022Does Moral Code Have a Moral Code? Probing Delphi's Moral Philosophycited by 31
  32. 2024Evaluating the Elementary Multilingual Capabilities of Large Language Models with MultiQcited by 27
  33. 2022The Authenticity Gap in Human Evaluationcited by 24
  34. 2023The Moral Machine Experiment on Large Language Modelscited by 24
  35. 2023InstructAlign: High-and-Low Resource Language Alignment via Continual Crosslingual Instruction Tuningcited by 21
  36. 2024High-Dimension Human Value Representation in Large Language Modelscited by 20
  37. 2018A Computational Model of Commonsense Moral Decision Makingcited by 13
  38. 2024Pre-Trained Language Models Represent Some Geographic Populations Better Than Otherscited by 6

50 most cited of 51 All 51 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 20; 2026: 26.

46 papers.

Sort:

By area:

  1. 2025Cultivating Pluralism In Algorithmic Monoculture: The Community Alignment Datasetcited by 32
  2. 2025Multiple LLM Agents Debate for Equitable Cultural Alignmentcited by 20
  3. 2025Synthetic Socratic Debates: Examining Persona Effects on Moral Decision and Persuasion Dynamicscited by 19
  4. 2025The Lock-in Hypothesis: Stagnation by Algorithmcited by 11
  5. 2025Discerning What Matters: A Multi-Dimensional Assessment of Moral Competence in LLMscited by 10
  6. 2025Stress-Testing Model Specs Reveals Character Differences among Language Modelscited by 10
  7. 2025Exploring Persona-dependent LLM Alignment for the Moral Machine Experimentcited by 9
  8. 2025LIVS: A Pluralistic Alignment Dataset for Inclusive Public Spacescited by 8
  9. 2025Agent-to-Agent Theory of Mind: Testing Interlocutor Awareness among Large Language Modelscited by 5
  10. 2025Language Specific Knowledge: Do Models Know Better in X than in English?cited by 4
  11. 2025Aligning Large Language Model Agents with Rational and Moral Preferences: A Supervised Fine-Tuning Approachcited by 4
  12. 2026Bias Beyond Borders: Political Ideology Evaluation and Steering in Multilingual LLMscited by 4
34 more
  1. 2025Multilingual LLMs Struggle to Link Orthography and Semantics in Bilingual Word Processingcited by 3
  2. 2025The Right to AIcited by 3
  3. 2025"Pull or Not to Pull?'': Investigating Moral Biases in Leading Large Language Models Across Ethical Dilemmascited by 3
  4. 2025ManagerBench: Evaluating the Safety-Pragmatism Trade-off in Autonomous LLMscited by 3
  5. 2025EVALUESTEER: Measuring Reward Model Steerability Towards Values and Preferencescited by 3
  6. 2026Evaluating Cooperation in LLM Social Groups through Elected Leadershipcited by 3
  7. 2025Dropouts in Confidence: Moral Uncertainty in Human-LLM Alignmentcited by 2
  8. 2026MOSAIC: Unveiling the Moral, Social and Individual Dimensions of Large Language Modelscited by 2
  9. 2026Moral Sensitivity in LLMs: A Tiered Evaluation of Contextual Bias via Behavioral Profiling and Mechanistic Interpretabilitycited by 2
  10. 2026Incoherent Values? Probing LLM Preferences Through Parametric Variationcited by 2
  11. 2025Reward Model Perspectives: Whose Opinions Do Reward Models Reward?cited by 1
  12. 2025Explainable Ethical Assessment on Human Behaviors by Generating Conflicting Social Normscited by 1
  13. 2026A Rising Tide Lifts All Boats: MTQE Rewards for Idioms Improve General Translation Qualitycited by 1
  14. 2026When Do Language Models Endorse Limitations on Human Rights Principles?cited by 1
  15. 2026Language Shapes Mental Health Evaluations in Large Language Modelscited by 1
  16. 2026Are Dilemmas and Conflicts in LLM Alignment Solvable? A View from Priority Graphcited by 1
  17. 2026GAIN: A Benchmark for Goal-Aligned Decision-Making of Large Language Models under Imperfect Normscited by 1
  18. 2026Conflicts Make Large Reasoning Models Vulnerable to Attackscited by 1
  19. 2026CulturALL: Benchmarking Multilingual and Multicultural Competence of LLMs on Grounded Taskscited by 1
  20. 2026Machine Behavior in Relational Moral Dilemmas: Moral Rightness, Predicted Human Behavior, and Model Decisionscited by 1
  21. 2026Every Act Has Its Price: Compressed Moral Composition in Frontier LLMscited by 1
  22. 2026D2VBench: Benchmarking Large Language Models with Value Dilemmas in Daily Scenarioscited by 1
  23. 2025Assessing Agentic Large Language Models in Multilingual National Biascited by 0
  24. 2026Learning to Diagnose and Correct Moral Errors: Beyond Shallow Heuristics in Moral Alignmentcited by 0
  25. 2026Building Interpretable Models for Moral Decision-Makingcited by 0
  26. 2026Towards Cross-lingual Values Judgment: A Consensus-Pluralism Perspectivecited by 0
  27. 2026ThaiSafetyBench: Assessing Language Model Safety in Thai Cultural Contextscited by 0
  28. 2026The Counterexample Game: Iterated Conceptual Analysis and Repair in Language Modelscited by 0
  29. 2026Training-Free Cultural Alignment of Large Language Models via Persona Disagreementcited by 0
  30. 2026Can LLMs Imagine Moral Alternatives Beyond Binary Dilemmas?cited by 0
  31. 2026The yes-no bias of large language models reflects answer order and wording, not shifts in moral judgmentcited by 0
  32. 2026MET: Theory-Grounded and Culture-Aware Multilingual Moral Reasoningcited by 0
  33. 2026How do LLMs Evaluate Perceived Moral Agency? Investigating Moral Decision-Making in Human-Artificial Agents Interactionscited by 0
  34. 2026Gender bias across LLMs is common and highly heterogeneouscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025One Model, Many Morals: Uncovering Cross-Linguistic Misalignments in Computational Moral Reasoning
  2. 2024The Moral Turing Test: Evaluating Human-LLM Alignment in Moral Decision-Making
  3. 2025Cultural Bias in Large Language Models: Evaluating AI Agents through Moral Questionnaires
  4. 2025Beyond Ethical Alignment: Evaluating LLMs as Artificial Moral Assistants
  5. 2024MoralBench: Moral Evaluation of LLMs
  6. 2024Decoding Multilingual Moral Preferences: Unveiling LLM's Biases Through the Moral Machine Experiment
  7. 2025Diverse Human Value Alignment for Large Language Models via Ethical Reasoning
  8. 2023Exploring the psychology of LLMs' Moral and Legal Reasoning