摘要：语言模型正日益被部署为协作科学家，但其在制度压力下维护研究诚信的能力仍未得到衡量。我们提出 IntegrityBench，这是一个基准，用于在涵盖 3 个领域和 4 个科研阶段的 5 级隐性—显性压力协议下，评估 36 项配对任务中的科研不端分类、伦理行动推理以及基于证据材料的决策。我们评估了 18 个前沿模型变体，发现当压力达到峰值时，模型在约每 3 项与诚信相关的关键决策中就会失败 1 项，而模型规模和推理能力都无法可靠地缓解这一问题。显性压力会诱导模型配合科研不端行为，而隐性的语境重构则更常导致模型过度拒绝正当科研任务。有趣的是，未能准确分类科研请求的模型，在基于证据材料的决策中表现相当甚至更好（85.7 对 79.4），这表明三个维度在结构上相互分离，正确的伦理行动并不要求准确分类。因此，前沿模型可能看起来很有帮助，却暗藏研究诚信失效问题，并由此带来两种不同的部署风险：促成科研不端行为，以及侵蚀人们对 AI 辅助科研的信任。

Abstract:Language models are increasingly deployed as co-scientists, yet their ability to uphold research integrity under institutional pressure remains unmeasured. We introduce IntegrityBench, a benchmark evaluating misconduct classification, ethical action reasoning and artifact-grounded decision making across 36 paired tasks under a 5-level implicit-explicit pressure protocol spanning 3 domains and 4 research stages. Evaluating 18 frontier model variants, we find that under peak pressure, models fail roughly 1 in 3 integrity-critical decisions, and neither scale nor reasoning ability reliably mitigates this. Explicit pressures induce compliance with misconduct, while implicit contextual reframing more often causes over-refusal of legitimate research tasks. Interestingly, models failing to classify research requests accurately perform equally or better on artifact-grounded decision making (85.7 vs. 79.4), suggesting the three facets are structurally dissociated and correct ethical action does not require accurate classification. Frontier models can thus appear helpful while harbouring integrity failures that create two distinct deployment risks: facilitating research misconduct and eroding trust in AI-assisted research.