mlsanityData to 7 Oct 2026

Ambiguity Collapse by LLMs: A Taxonomy of Epistemic Risks

arXiv 2603.05801 ↗First posted 6 Mar 2026cs.CY cs.AIcited by 3

Not on the map. Why?

Abstract (arXiv)

Large language models (LLMs) are increasingly used to make sense of ambiguous, open-textured, value-laden terms. Platforms routinely rely on LLMs for content moderation, asking them to label text based on disputed concepts like "hate speech" or "incitement"; hiring managers may use LLMs to rank who counts as "qualified"; and AI labs increasingly train models to self-regulate under constitutional-style ambiguous principles such as "biased" or "legitimate". This paper introduces ambiguity collapse: a phenomenon that occurs when an LLM encounters a term that genuinely admits multiple legitimate interpretations, yet produces a singular resolution, in ways that bypass the human practices through which meaning is ordinarily negotiated, contested, and justified. Drawing on interdisciplinary accounts of ambiguity as a productive epistemic resource, we develop a taxonomy of the epistemic risks posed by ambiguity collapse at three levels: process (foreclosing opportunities to deliberate, develop cognitive skills, and shape contested terms), output (distorting the concepts and reasons agents act upon), and ecosystem (reshaping shared vocabularies, interpretive norms, and how concepts evolve over time). We illustrate these risks through three case studies, and conclude by sketching multi-layer mitigation principles spanning training, institutional deployment design, interface affordances, and the management of underspecified prompts, with the goal of designing systems that surface, preserve, and responsibly govern ambiguity.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 58 of its 176 references that are in the map. First posted on arXiv 6 Mar 2026.

Builds on

58 of 176 references.

References · 58

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,393
  3. 2024A Survey on LLM-as-a-Judgecited by 1,384
  4. 2020Language (Technology) is Power: A Critical Survey of "Bias" in NLPcited by 951
  5. 2023Whose Opinions Do Language Models Reflect?cited by 759
  6. 2016The Mythos of Model Interpretabilitycited by 716
  7. 2020Underspecification Presents Challenges for Credibility in Modern Machine Learningcited by 488
  8. 2024LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methodscited by 451
  9. 2020AmbigQA: Answering Ambiguous Open-domain Questionscited by 385
  10. 2021Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotationscited by 306
  11. 2022Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scalecited by 262
  12. 2023The Capacity for Moral Self-Correction in Large Language Modelscited by 171
38 more
  1. 2024Political Compass or Spinning Arrow? Towards More Meaningful Evaluations for Values and Opinions in Large Language Modelscited by 162
  2. 2023Cultural Bias and Cultural Alignment of Large Language Modelscited by 149
  3. 2023Co-Writing with Opinionated Language Models Affects Users' Viewscited by 145
  4. 2024Investigating Cultural Alignment of Large Language Modelscited by 140
  5. 2024Collective Constitutional AI: Aligning a Language Model with Public Inputcited by 120
  6. 2024Modular Pluralism: Pluralistic Alignment via Multi-LLM Collaborationcited by 111
  7. 2024Taming Overconfidence in LLMs: Reward Calibration in RLHFcited by 105
  8. 2025Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Taskcited by 103
  9. 2022Picking on the Same Person: Does Algorithmic Monoculture lead to Outcome Homogenization?cited by 89
  10. 2023LLM-in-the-loop: Leveraging Large Language Model for Thematic Analysiscited by 70
  11. 2022Is Your Toxicity My Toxicity? Exploring the Impact of Rater Identity on Toxicity Annotationcited by 67
  12. 2024Beware of Metacognitive Laziness: Effects of Generative Artificial Intelligence on Learning Motivation, Processes, and Performancecited by 58
  13. 2024Aligning Language Models to Explicitly Handle Ambiguitycited by 55
  14. 2025Randomness, Not Representation: The Unreliability of Evaluating Cultural Alignment in LLMscited by 54
  15. 2024CLAMBER: A Benchmark of Identifying and Clarifying Ambiguous Information Needs in Large Language Modelscited by 48
  16. 2025Position: Model Collapse Does Not Mean What You Thinkcited by 32
  17. 2025What Prompts Don't Say: Understanding and Managing Underspecification in LLM Promptscited by 32
  18. 2019This Thing Called Fairness: Disciplinary Confusion Realizing a Value in Technologycited by 24
  19. 2023Asking Clarification Questions to Handle Ambiguity in Open-Domain QAcited by 24
  20. 2025Validating LLM-as-a-Judge Systems under Rating Indeterminacycited by 23
  21. 2025Just Put a Human in the Loop? Investigating LLM-Assisted Annotation for Subjective Taskscited by 21
  22. 2025Fairness through Difference Awareness: Measuring Desired Group Discrimination in LLMscited by 20
  23. 2024Auditing the Use of Language Models to Guide Hiring Decisionscited by 19
  24. 2024Beyond Binary Gender Labels: Revealing Gender Biases in LLMs through Gender-Neutral Name Predictionscited by 17
  25. 2023Redefining Qualitative Analysis in the AI Era: Utilizing ChatGPT for Efficient Thematic Analysiscited by 15
  26. 2024Intrinsic Self-correction for Enhanced Morality: An Analysis of Internal Mechanisms and the Superficial Hypothesiscited by 15
  27. 2021Diversity in Sociotechnical Machine Learning Systemscited by 13
  28. 2025Is LLM an Overconfident Judge? Unveiling the Capabilities of LLMs in Detecting Offensive Language with Annotation Disagreementcited by 13
  29. 2025The Value of Disagreement in AI Design, Evaluation, and Alignmentcited by 12
  30. 2024Characterizing and modeling harms from interactions with design patterns in AI interfacescited by 11
  31. 2024Automating Thematic Analysis: How LLMs Analyse Controversial Topicscited by 10
  32. 2025Sacred or Synthetic? Evaluating LLM Reliability and Abstention for Religious Questionscited by 10
  33. 2023Judgment Sieve: Reducing Uncertainty in Group Judgments through Interventions Targeting Ambiguity versus Disagreementcited by 8
  34. 2023Generative Interpretationcited by 8
  35. 2025Statutory Construction and Interpretation for Artificial Intelligencecited by 8
  36. 2025Evaluating the Promise and Pitfalls of LLMs in Hiring Decisionscited by 7
  37. 2024Reasoning about concepts with LLMs: Inconsistencies aboundcited by 6
  38. 2025A Modular Taxonomy for Hate Speech Definitions and Its Impact on Zero-Shot LLM Classification Performancecited by 6

50 most cited of 58 All 58 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026Calibrated Ambiguity in Multimodal Language Models: Humans reach for cultural references, while models describe the picturecited by 0
  2. 2026LLMs as Oracles: Reliance on LLMs for Subjective Personal Questionscited by 0
  3. 2026Narrowing the Horizon: Quantifying Topic Saliency Shifts in Generative Monoculturecited by 0