arXiv cs.AI· Kainan Zhou, Chuhong Xu, Gangzhen Qian, Zhaoyi Li·· 6 天前AI 评分18
面向生成式搜索的 Claim-Gated 来源风险审计
Claim-Gated Source-Risk Auditing for Generative Search
AI 导读
研究提出一种针对查询-来源-答案三元组的 claim-gated 审计规范，仅当关系证据、答案采纳、重要性与披露四项均被观察到时才判定遗漏成立，证据不完整则保持未决而非视为独立。配套参考检查器在合成测试集上复现全部 81 种三态谓词组合，并拒绝 192 条故意构造的畸形记录。作者强调这些是有限契约一致性结果，非检测器准确率，语义效度与部署收益仍需独立标注、留出评估与配对效用测试验证。
正文 · 原文
Abstract:A generative search answer can cite a supported passage yet omit a source relationship that changes its interpretation. We specify a claim-gated audit of the query-source-answer tuple. An omission is resolved only when relationship evidence, answer adoption, materiality, and disclosure are all observed; incomplete evidence remains unresolved rather than being treated as independence. The specification separates this endpoint from citation support and review priority, and binds decisions to versioned evidence spans. A reference checker makes the record contract executable. On an exhaustive synthetic suite, it reproduces all 81 three-state predicate combinations and rejects 192 deliberately malformed records. Common-guard baselines and predicate ablations isolate endpoint logic from missing-evidence handling, while controlled transitions check support separation and evidence removal. These are finite contract-conformance results, not detector accuracy or evidence of improved user outcomes. We define the independent annotation, held-out evaluation, and paired utility tests still required to establish semantic validity and deployment benefit.
| Comments: | 7 pages, 2 figures, 4 tables. Accepted at IEEE AI2A 2026. Author accepted manuscript. Reproducibility artifact: this https URL |
| Subjects: | Artificial Intelligence (cs.AI); Computational Complexity (cs.CC); Computational Engineering, Finance, and Science (cs.CE); Emerging Technologies (cs.ET); Information Retrieval (cs.IR) |
| Cite as: | arXiv:2609.29145 [cs.AI] |
| (or arXiv:2609.29145v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2609.29145 arXiv-issued DOI via DataCite |
Submission history
From: Kainan Zhou [view email]
[v1]
Thu, 24 Sep 2026 07:22:14 UTC (267 KB)
[v2]
Fri, 9 Oct 2026 05:01:42 UTC (267 KB)
来源：arXiv cs.AI · arxiv.org