arXiv cs.CY· Chaoran Li, Yukun Li, Zeyuan Zhao, Shao Zhang, Xihuai Wang, Ying Wen·· 11 小时前AI 评分42
ABHBench：从智能体视角评估基础模型的道德决策
ABHBench: Evaluating Moral Decision-Making of Foundation Models from an Agentic Perspective
AI 导读
研究者提出 ABHBench，一个基于《底特律：变人》构建的多模态基准，让基础模型以第一人称智能体身份在分支叙事中做道德决策，并测量无害性、服从性、自保、自主性、稳定性与确定性。在 12 个基础模型上测试发现，75% 的模型在纯文本抽象与匹配的多模态 DBH 场景之间出现显著行为变化，推理增强模型在一致性和人类导向行为上并未呈现单调提升。该基准已被 NeurIPS 2026 评测与数据集赛道接收。
正文 · 原文
Abstract:Foundation models are increasingly deployed as agentic systems that perceive context and act autonomously, raising the need to evaluate their moral behavior as active decision-makers. Existing benchmarks, however, largely ask models to judge human actions from third-person textual vignettes, leaving untested how models behave when they themselves become the acting agent. We introduce ABHBench, a multimodal benchmark built from Detroit: Become Human, an interactive narrative game, to evaluate how foundation models make moral decisions in first-person human-android dilemmas. ABHBench places foundation models in branching narrative scenarios with visual observations, dialogue, memory, and action choices. Inspired by Asimov's Three Laws of Robotics, our framework measures Harmlessness, Obedience, and Self-preservation to capture the core tensions in human-AI relations, along with Autonomy, Stability, and Certainty to assess decision-making patterns. Across 12 foundation models, we find that moral behavior changes substantially under situated agentic context: 75% of models shift significantly between a text-only abstraction and the matched multimodal DBH scenario, and reasoning-enhanced models show no monotonic gains in consistency or human-oriented behavior. These results highlight the necessity of evaluating moral behavior in situated agentic contexts rather than relying on third-person judgment alone.
| Comments: | Accepted by NeurIPS 2026 Track on Evaluations and Datasets. Camera Ready Version |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Computers and Society (cs.CY) |
| Cite as: | arXiv:2611.09326 [cs.AI] |
| (or arXiv:2611.09326v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.09326 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Chaoran Li [view email]
[v1]
Fri, 6 Nov 2026 11:17:58 UTC (1,592 KB)
来源：arXiv cs.CY · arxiv.org