📄 中文摘要
随着人工智能代理在经济活动中扮演越来越重要的角色（如执行交易、管理预算、谈判合同和生成子代理），当前的框架将这种经济代理的权限限制在与操作稳健性无关的能力基准上。提出了一种名为理解门控代理经济（CGAE）的正式架构，其中代理的经济权限由经过验证的理解函数上限，该函数源自对抗性稳健性审计。门控机制在三个正交的稳健性维度上运行：约束合规性（通过CDCT测量）、认知完整性（通过DDFT测量）和行为一致性（通过AGT测量），而内在的幻觉率则作为一个交叉切割的指标。
理解门控的代理经济：以稳健性为首的人工智能经济代理架构
📄 English Summary
The Comprehension-Gated Agent Economy: A Robustness-First Architecture for AI Economic Agency
The Comprehension-Gated Agent Economy (CGAE) proposes a formal architecture that limits the economic permissions of AI agents based on a verified comprehension function derived from adversarial robustness audits. This approach addresses the increasing role of AI agents in economic activities such as executing trades, managing budgets, negotiating contracts, and spawning sub-agents. Current frameworks often tie economic agency to capability benchmarks that lack correlation with operational robustness. The gating mechanism operates across three orthogonal dimensions of robustness: constraint compliance (measured by CDCT), epistemic integrity (measured by DDFT), and behavioral alignment (measured by AGT), with intrinsic hallucination rates serving as a cross-cutting metric. This architecture aims to enhance the reliability and effectiveness of AI agents in economic contexts.