arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.14606cs.MAcs.AIcs.CE
收藏

迈向选择作为权力:自主代理中决策权威的边界

Towards Selection as Power: Bounding Decision Authority in Autonomous Agents

AI Lab, Grupo Santander Madrid, Spain(AI实验室,西班牙 Grupo Santander 马德里)

Jose Manuel de la Chica Rodriguez, Juan Manuel Vera Díaz

AI总结:

本文提出一种治理架构,通过机械机制限制自主代理的选择权,以防止确定性结果并提升可审计性,重新定义治理为受限制的因果权力。

摘要:

自主代理系统越来越多地部署在受监管、高风险的领域中,其中决策可能是不可逆的,并且受到制度限制。现有的安全方法强调对齐、可解释性或行动层面的过滤。我们认为这些机制是必要的但不够,因为它们不直接控制选择权:决定哪些选项被生成、呈现和框定的权威。我们提出了一种治理架构,将认知、选择和行动分为不同的领域,并将自主性建模为主权的向量。认知自主性保持不受限制,而选择和行动自主性则通过机械强制的原始机制在代理的优化空间之外进行限制。该架构整合了外部候选生成(CEFL)、受控的缩减器、提交-揭示熵隔离、理由验证以及故障响亮的电路断路器。我们评估了该系统在多个受监管的金融场景中,针对方差操纵、阈值游戏、框架偏差、顺序效应和熵探测的对抗性压力。指标量化了选择集中度、叙述多样性、治理激活成本和失败可见性。结果显示,机械选择治理是可实现、可审计的,并且能防止确定性结果捕获,同时保持推理能力。尽管概率集中度仍然存在,但该架构相对于传统标量管道在可测量地限制选择权方面具有优势。这项工作将治理重新定义为受限制的因果权力,而不是内部意图对齐,为部署自主代理提供了基础,其中沉默失败是不可接受的。

英文摘要:

Autonomous agentic systems are increasingly deployed in regulated, high-stakes domains where decisions may be irreversible and institutionally constrained. Existing safety approaches emphasize alignment, interpretability, or action-level filtering. We argue that these mechanisms are necessary but insufficient because they do not directly govern selection power: the authority to determine which options are generated, surfaced, and framed for decision. We propose a governance architecture that separates cognition, selection, and action into distinct domains and models autonomy as a vector of sovereignty. Cognitive autonomy remains unconstrained, while selection and action autonomy are bounded through mechanically enforced primitives operating outside the agent's optimization space. The architecture integrates external candidate generation (CEFL), a governed reducer, commit-reveal entropy isolation, rationale validation, and fail-loud circuit breakers. We evaluate the system across multiple regulated financial scenarios under adversarial stress targeting variance manipulation, threshold gaming, framing skew, ordering effects, and entropy probing. Metrics quantify selection concentration, narrative diversity, governance activation cost, and failure visibility. Results show that mechanical selection governance is implementable, auditable, and prevents deterministic outcome capture while preserving reasoning capacity. Although probabilistic concentration remains, the architecture measurably bounds selection authority relative to conventional scalar pipelines. This work reframes governance as bounded causal power rather than internal intent alignment, offering a foundation for deploying autonomous agents where silent failure is unacceptable.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑