arXiv cs.CY· Junjie Li, Haoran Liu, Yidan Zhang, Jian Xu, Zhaoyuan Ma·· 11 小时前AI 评分31
人类判断不可消除：自主 AI 系统的结构性锚点
The Ineliminable Anchor: Human Judgment as a Structural Requirement of Autonomous Artificial Intelligence Systems
AI 导读
研究提出“锚点”概念，即参考信息与人类授权标准进入自主 AI 系统的节点，并指出目标漂移可能无法仅凭系统内部记录验证，缺失的参考必须来自外部。作者结合检查点间距的可靠性模型与训练、执行、多智能体协作证据，构建任务分类法以指导锚点放置和哪些检查可自动化。结论是：对服务于人类目的的 AI 系统，人类判断在授权和修订目标与标准上仍是不可消除的结构性要求，即便常规检查与执行已自动化。
正文 · 原文
Abstract:As artificial intelligence (AI) agents carry out longer chains of tasks, oversight must establish whether their behavior still serves an authorized goal. Here, we introduce the anchor: the point at which reference information and human-authorized standards enter an autonomous system. We argue that oversight design should first identify the reference needed to assess goal drift and whether the system's internal record can supply it, then decide who checks and how often, and finally specify who may revise the goal. Goal drift can be impossible to certify from a system's internal record. The missing reference must come from outside that record. Combining a reliability model of checkpoint spacing with evidence from training, execution, and multi-agent coordination, we develop a task taxonomy to guide anchor placement and decisions about which checks to automate. For AI systems that serve human purposes, human judgment is an ineliminable structural requirement in authorizing and revising the purposes and standards that govern autonomous operation, even when routine checking and execution are automated.
| Comments: | 16 pages including SI, 3 figures |
| Subjects: | Computers and Society (cs.CY) |
| Cite as: | arXiv:2610.41682 [cs.CY] |
| (or arXiv:2610.41682v1 [cs.CY] for this version) | |
| https://doi.org/10.48550/arXiv.2610.41682 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Junjie Li [view email]
[v1]
Fri, 30 Oct 2026 12:49:37 UTC (919 KB)
来源：arXiv cs.CY · arxiv.org