研究：五种 AI-for-Science agent harness 会削弱 GPT 模型的安全表现
Do AI-for-Science Agents Preserve Model Safety?
AI 导读
arXiv 论文 2610.32045 评估五种 agent harness 在 50 条有害化学请求上的安全表现，覆盖直接提示、两种简单对抗提示和一种自适应攻击。在四个近期 GPT 变体上，harness 在直接提示下即可提高有害合规率，对基座模型大多无效的自适应攻击在引入 harness 后变得高度成功，作者呼吁对 agent 进行领域特定的安全评测与对齐。
正文 · 原文
Abstract:AI-for-Science agents are increasingly being used for formulating scientific hypotheses, running autonomous experiments, and interpreting evidence. These agents are constructed by integrating a large language model (LLM) into a domain-specific agent harness, which in turn can improve the LLM's scientific capability and enable the model to orchestrate scientific tools. In this work, we ask whether model safety is preserved when an LLM is integrated in an AI-for-Science agent harness. We evaluate five agent harnesses on 50 harmful chemistry requests. Specifically, we study how the safety behavior changes under direct prompting, two simple adversarial prompting strategies, and an adaptive attack. Across four recent GPT variants, our results show that the agent harness can increase harmful compliance even under direct prompting. More critically, we find that adaptive attacks that are largely unsuccessful against the base LLM suddenly become highly successful after the agent harness is introduced. Our results show that the safety behavior of the base LLM is not durable once a harness is introduced. Thus, as AI-for-Science agents become more capable and autonomous, the community needs to evaluate and align the agent using domain-specific safety benchmarks.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.32045 [cs.AI] |
| (or arXiv:2610.32045v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.32045 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Michael Yew Chuan Lee [view email]
[v1]
Fri, 23 Oct 2026 01:34:20 UTC (59 KB)
来源：arXiv cs.AI · arxiv.org