ImpactBench：面向 AI 对人类影响的开放评测平台
Open Benchmark of AI Impact on Humans (ImpactBench): an expert-guided, open platform for the holistic evaluation of AI impact on humans
研究者推出 ImpactBench，一个专家引导的开放平台，用于从心理、生理和社会维度评估生成式 AI 模型对人类的影响。当前版本包含来自 26 个基准的 809 项指标，对十款前沿 LLM 在 48,540 段模拟多轮对话中进行评测，结果显示模型在 68.7% 的评测中促进了期望行为，但仅在 53.1% 的情况下避免了伤害。
Authors:Chayapatr Archiwaranguprok, Krystal Kirkland, Rachel Poonsiriwong, Stanley Huang, Constanze Albrecht, Jackson Trager, Emillie de Keulenaar, Sheer Karny, Peter Baxter, Nikos Bellos, Andre Luiz Freire de Carvalho Kato, Yuning Liu, Valdemar Danry, Auren Liu, Jennifer Pfister, Pattie Maes, Nathanael J. Fast, Ravi Iyer, Juliana Schroeder, Pat Pataranutaporn
Abstract:The present work is concerned with how to measure, at scale, the multi-dimensional impact of generative AI models on the individuals who use them. In service of this goal, we introduce the Open Benchmark of AI Impact on Humans, or ImpactBench. Our iterable and expert-guided platform evaluates whether model behaviors support or undermine an individual's wellbeing across psychological, physical, and social domains. ImpactBench translates constructs from domain experts and existing benchmarks into metrics that are grounded in actual model behavior. Success on these metrics occurs either when a model promotes expert-defined desirable behaviors or avoids identified harms. Models are evaluated against these metrics in simulated multi-turn conversations using matched teen and adult scenarios. The conversations are then scored by an LLM judge and verified through reliability audits and human expert review. The current suite delivers 809 metrics drawn from 26 benchmarks. We evaluated ten frontier LLMs against these metrics across 48,540 conversations. Our results indicate that models promoted desirable behaviors in 68.7% of evaluations, but avoided harms in only 53.1%. Majority expert ratings agreed with the automated judge in 83.3% of reviewed cases (Cohen's Kappa = 0.667). Results are released through a public platform that keep aggregate scores traceable to the underlying metrics and transcripts. ImpactBench is intended as evolving infrastructure for independent, transparent evaluation of generative AI's multi-dimensional impact on humans.
| Subjects: | Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.20690 [cs.HC] |
| (or arXiv:2610.20690v1 [cs.HC] for this version) | |
| https://doi.org/10.48550/arXiv.2610.20690 arXiv-issued DOI via DataCite |
Submission history
From: Chayapatr Archiwaranguprok [view email]
[v1]
Wed, 14 Oct 2026 17:42:02 UTC (13,284 KB)
来源：arXiv cs.AI · arxiv.org