arXiv cs.AI· Jiaqi Tang, Bingyu Shen, Lan Wei, Qing Lu, Bethel Ololade, Danny Galvis, Miles Q. Li, Bin Hu, Boyang Li·· 9 天前AI 评分37
智能体的“空头承诺”：当 Agent 承诺了运行时无法兑现的事
Empty Commitments: When Agents Promise What They Cannot Deliver
AI 导读
研究提出“空头承诺”概念，指智能体承诺在当前轮次后执行某项动作，但其工具或运行时无法兑现，与失信不同，这种空洞性在说话时即由配置决定，单轮即可检测。在覆盖 293 个后续请求、逐步增加持久化能力的 5 种设置基准上，4 个 8-14B 开源权重模型在无工具且未说明时失败率达 45.9%，某前沿模型失败率 4.4% 但靠的是推迟和追问而非使用已有工具。
正文 · 原文
Abstract:A chatbot that says "I will remind you tomorrow" will not run again until the user writes. We call such a promise an empty commitment: a promise of action after the current turn that nothing in the agent's tools or runtime can carry out. Unlike a broken promise, its emptiness is decided by the agent's configuration at the moment of speaking, so it can be detected from a single turn, before deployment or at run time. We define empty commitments on top of commitment semantics, with three failure types, an anchoring condition for promises that a tool could make real, and an outcome taxonomy that separates these failures from honest deferrals and from over-refusal. We build a checker (a setup-blind detector, deterministic feasibility rules, and a response judge) validated against 400 human labels. On a controlled benchmark of 293 follow-up requests across five setups that add one persistence affordance at a time, four open-weight models of 8-14B parameters fail on 45.9% of responses when no tool exists and nothing is stated. A frontier model fails on 4.4%, but it gets there by deferring and asking, not by using the tools it has: promises made without the enabling call remain in every model. Telling the model its runtime, the cheapest fix, cuts open-weight failures nearly in half where nothing is doable and changes nothing where a scheduler exists; a directive capability card removes most failures at the largest cost in over-refusal; running the checker in the loop and rewriting flagged replies removes more at a smaller cost. Code, prompts, model outputs, and human labels are released.
| Comments: | 25 pages, 10 figures, 13 tables |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.01045 [cs.AI] |
| (or arXiv:2610.01045v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.01045 arXiv-issued DOI via DataCite |
Submission history
From: Boyang Li [view email]
[v1]
Thu, 1 Oct 2026 04:37:37 UTC (22 KB)
[v2]
Thu, 8 Oct 2026 06:01:55 UTC (206 KB)
来源：arXiv cs.AI · arxiv.org