# Accommodation and Epistemic Vigilance: A Pragmatic Account of Why LLMs Fail to Challenge Harmful Beliefs

arXiv: https://arxiv.org/abs/2601.04435 · HTML: https://mlsanity.com/paper/2601.04435/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 44 references that are in the map.
- First posted on arXiv 7 Jan 2026.

## Builds on (most cited first, 14 of 14)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Navigating the Grey Area: How Expressions of Uncertainty and Overconfidence Affect Language Models](https://mlsanity.com/paper/2302.13439/) (2023)
- [In conversation with Artificial Intelligence: aligning language models with human values](https://mlsanity.com/paper/2209.00731/) (2022)
- [Grounding Gaps in Language Model Generations](https://mlsanity.com/paper/2311.09144/) (2023)
- [Confabulation: The Surprising Value of Large Language Model Hallucinations](https://mlsanity.com/paper/2406.04175/) (2024)
- [From tools to thieves: Measuring and understanding public perceptions of AI through crowdsourced metaphors](https://mlsanity.com/paper/2501.18045/) (2025)
- [Large Models of What? Mistaking Engineering Achievements for Human Linguistic Agency](https://mlsanity.com/paper/2407.08790/) (2024)
- [Comparing human and LLM politeness strategies in free production](https://mlsanity.com/paper/2506.09391/) (2025)
- [SAGE-Eval: Evaluating LLMs for Systematic Generalizations of Safety Facts](https://mlsanity.com/paper/2505.21828/) (2025)
- [Martingale Score: An Unsupervised Metric for Bayesian Rationality in LLM Reasoning](https://mlsanity.com/paper/2512.02914/) (2025)
- [Is It JUST Semantics? A Case Study of Discourse Particle Understanding in LLMs](https://mlsanity.com/paper/2506.04534/) (2025)
- [From Fact to Judgment: Investigating the Impact of Task Framing on LLM Conviction in Dialogue Systems](https://mlsanity.com/paper/2511.10871/) (2025)
- [Hey, wait a minute: on at-issue sensitivity in Language Models](https://mlsanity.com/paper/2510.12740/) (2025)
- [Epistemic Fragility in Large Language Models: Prompt Framing Systematically Modulates Misinformation Correction](https://mlsanity.com/paper/2511.22746/) (2025)

## Built on by (most cited first, 11 of 11)

- [BeliefShift: Benchmarking Temporal Belief Consistency and Opinion Drift in LLM Agents](https://mlsanity.com/paper/2603.23848/) (2026)
- [AI Assistance Reduces Persistence and Hurts Independent Performance](https://mlsanity.com/paper/2604.04721/) (2026)
- [SWAY: A Counterfactual Computational Linguistic Approach to Measuring and Mitigating Sycophancy](https://mlsanity.com/paper/2604.02423/) (2026)
- [Personalization Increases Affective Alignment but Has Role-Dependent Effects on Epistemic Independence in LLMs](https://mlsanity.com/paper/2603.00024/) (2026)
- [Verbalizing LLMs' assumptions to explain and control sycophancy](https://mlsanity.com/paper/2604.03058/) (2026)
- [Sycophancy Undermines Epistemic Vigilance in Cooperative Vision-Language Tasks](https://mlsanity.com/paper/2607.29585/) (2026)
- [Don't `Well, Actually' Me Unless You Know What You're Talking About: Weak Presupposition Verification Degrades General QA Performance](https://mlsanity.com/paper/2608.06539/) (2026)
- [The Civilization Framework: Sovereign-Anchored Communication Between Personal Multi-Agent Systems](https://mlsanity.com/paper/2609.03425/) (2026)
- [Mitigating Social Sycophancy via Pluralistic Preference Optimization](https://mlsanity.com/paper/2610.02568/) (2026)
- [TrustMI: Causally controlling how assistants trust their users](https://mlsanity.com/paper/2610.06064/) (2026)
- [Pushback in a Haystack: Language Models Respond to Factual Errors Ambiguously and Ineffectively](https://mlsanity.com/paper/2610.17703/) (2026)

## Abstract (arXiv)

Large language models (LLMs) frequently fail to challenge users' harmful beliefs in domains ranging from medical advice to social reasoning. We argue that these failures can be understood and addressed pragmatically as consequences of LLMs defaulting to accommodating users' assumptions and exhibiting insufficient epistemic vigilance. We show that social and linguistic factors known to influence accommodation in humans (at-issueness, linguistic encoding, and source reliability) similarly affect accommodation in LLMs, explaining performance differences across three safety benchmarks that test models' ability to challenge harmful beliefs, spanning misinformation (Cancer-Myth, SAGE-Eval) and sycophancy (ELEPHANT). We further show that simple pragmatic interventions, such as adding the phrase "wait a minute", significantly improve performance on these benchmarks while preserving low false-positive rates. Our results highlight the importance of considering pragmatics for evaluating LLM behavior and improving LLM safety.