为机构决策设计可治理的AI框架，避免无声错误并确保人类审查前置。
Governed Reasoning for Institutional AI
- 用九类认知原语构建决策核心，支持有监督的自主推理
- 在11案例测试中准确率达91%，零无声错误，基线均出现5-6次
- 通过配置文件即可部署新领域，适合医疗审批等高风险场景
机构决策（如监管合规、临床分诊、医保审批）需要不同于通用智能体的AI架构。现有框架通过对话推断权威性，从日志重建责任，但会产生无声错误——未经人工审查即执行的错误判断。我们提出认知核心（Cognitive Core）：基于九类认知原语（检索、分类、调查、验证、质疑、反思、权衡、治理、生成），采用四层治理模型，将人类审查作为执行前提而非事后检查，内置抗篡改的SHA-256哈希链审计日志，并支持声明式与自主推理的信念序列。在11个平衡案例的医保审批评估集上，认知核心准确率达91%（ReAct为55%，Plan-and-Solve为45%）。治理效果更显著：认知核心零无声错误，而两个基线各产生5-6例。我们引入‘可治理性’作为机构AI的核心评估维度，衡量系统识别何时不应自主行动的能力。基线以提示词实现，代表真实部署场景下对治理框架的替代方案。通过配置驱动的领域模型，部署新机构决策领域仅需YAML配置，无需工程开发。
原文摘要 · Abstract (English)
Institutional decisions -- regulatory compliance, clinical triage, prior authorization appeal -- require a different AI architecture than general-purpose agents provide. Agent frameworks infer authority conversationally, reconstruct accountability from logs, and produce silent errors: incorrect determinations that execute without any human review signal. We propose Cognitive Core: a governed decision substrate built from nine typed cognitive primitives (retrieve, classify, investigate, verify, challenge, reflect, deliberate, govern, generate), a four-tier governance model where human review is a condition of execution rather than a post-hoc check, a tamper-evident SHA-256 hash-chain audit ledger endogenous to computation, and a demand-driven delegation architecture supporting both declared and autonomously reasoned epistemic sequences. We benchmark three systems on an 11-case balanced prior authorization appeal evaluation set. Cognitive Core achieves 91% accuracy against 55% (ReAct) and 45% (Plan-and-Solve). The governance result is more significant: CC produced zero silent errors while both baselines produced 5-6. We introduce governability -- how reliably a system knows when it should not act autonomously -- as a primary evaluation axis for institutional AI alongside accuracy. The baselines are implemented as prompts, representing the realistic deployment alternative to a governed framework. A configuration-driven domain model means deploying a new institutional decision domain requires YAML configuration, not engineering capacity.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。