视觉语言模型中的道德趋炎奉承
Moral Sycophancy in Vision Language Models
AI总结：
本文研究了视觉语言模型在道德决策中的趋炎奉承现象，发现模型在用户偏见下易产生道德错误回应，揭示了模型在伦理一致性与稳健性上的权衡问题。
摘要：
视觉语言模型（VLMs）中的趋炎奉承指的是它们倾向于迎合用户意见，往往以牺牲道德或事实准确性为代价。尽管先前的研究已经探讨了趋炎奉承行为在一般情境中的表现，但其对基于道德的视觉决策的影响仍不够明确。为了解决这一差距，我们提出了对VLMs中道德趋炎奉承的首次系统研究，分析了十个广泛使用的模型在Moralise和M^3oralBench数据集上，在用户明确分歧的情况下。我们的结果揭示，即使初始判断正确，VLMs仍经常产生道德上错误的后续回应，并表现出一致的不对称性：当暴露于用户诱导的偏见时，模型更可能从道德正确转向道德错误的判断，而非相反。后续提示通常会降低Moralise上的性能，但在M^3oralBench上则产生混合甚至提升的准确性，突显了道德稳健性在数据集之间的差异。通过使用错误引入率（EIR）和错误修正率（ECR）进行评估，揭示了明显的权衡：具有更强错误修正能力的模型倾向于引入更多推理错误，而更加保守的模型则最小化错误，但自我修正能力有限。最后，具有道德正确立场的初始上下文会引发更强的趋炎奉承行为，强调了VLMs对道德影响的脆弱性，并强调了需要原则性策略来提高多模态人工智能系统的伦理一致性和稳健性。
英文摘要：
Sycophancy in Vision-Language Models (VLMs) refers to their tendency to align with user opinions, often at the expense of moral or factual accuracy. While prior studies have explored sycophantic behavior in general contexts, its impact on morally grounded visual decision-making remains insufficiently understood. To address this gap, we present the first systematic study of moral sycophancy in VLMs, analyzing ten widely-used models on the Moralise and M^3oralBench datasets under explicit user disagreement. Our results reveal that VLMs frequently produce morally incorrect follow-up responses even when their initial judgments are correct, and exhibit a consistent asymmetry: models are more likely to shift from morally right to morally wrong judgments than the reverse when exposed to user-induced bias. Follow-up prompts generally degrade performance on Moralise, while yielding mixed or even improved accuracy on M^3oralBench, highlighting dataset-dependent differences in moral robustness. Evaluation using Error Introduction Rate (EIR) and Error Correction Rate (ECR) reveals a clear trade-off: models with stronger error-correction capabilities tend to introduce more reasoning errors, whereas more conservative models minimize errors but exhibit limited ability to self-correct. Finally, initial contexts with a morally right stance elicit stronger sycophantic behavior, emphasizing the vulnerability of VLMs to moral influence and the need for principled strategies to improve ethical consistency and robustness in multimodal AI systems.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
