论文提出长效个人 AI 智能体的激励层框架与运行时架构
Incentive-Responsive Agents: A Principal-Agent Framework, Metric Suite, and Runtime Architecture for Long-Lived Personal AI Agents
一篇把经济学主代理模型套在 AI 智能体上的论文，实测提示词里写激励没用，改运行时资源分配后最难任务成功率从 0.60 拉到 1.00。做智能体调度、想给 Agent 上的算力和权限定规则的开发者值得看，代码和仿真都开源了。
arXiv 论文《Incentive-Responsive Agents》主张长期运行的 AI 智能体缺的不是能力而是激励层。作者把激励形式化为运行时契约，覆盖模型档位、测试时算力、记忆预算、工具权限、任务优先级等运行点，并给出非消费不变性、持续身份折扣定理、激励-能力汇率前沿和 Goodhart 上界四项理论结果。
实验用两个模型和 60 个可验证任务跑了 240 次：提示词层面的激励措辞无显著效果（p = 0.62），而运行时资源分配把最难任务组的成功率从 0.60 提升到 1.00（p = 0.008），代价是 4.8 倍价格加权成本，汇率定理预测盈亏平衡基线为 0.21。另有 20 个随机种子、3000 个任务的仿真显示资源 ROI 从 7.09 升到 16.67，作弊率从 0.333 降到 0.115。
这篇论文对做智能体系统的开发者有直接参考价值：它把主-代理经济学（隐藏努力、可契约运行点、验证成本）引入 LLM 智能体调度，比单纯堆提示词或换更大模型更可度量。代码、仿真器和全部结果已开源，实际在线部署尚未完成。
Incentive-Responsive Agents: A Principal-Agent Framework, Metric Suite, and Runtime Architecture for Long-Lived Personal AI Agents
Personal AI agents now persist: they hold accounts, mailboxes, long-term memory, and open-ended task streams, yet the dominant way to improve them remains capability-side: prompt engineering, larger models, longer thinking, and retrieval. We argue that long-lived agents are missing an incentive layer. We formalize an incentive not as text in a prompt but as a runtime contract over the agent's operating point (model tier, test-time compute, memory budget, tool authority, task priority, verification, and compensation) that the agent's policy must causally consume. We develop a principal-agent model with hidden effort, a contractible operating point, and costly verification. We derive (i) a non-consumption invariance principle separating genuine incentives from prompt-level persuasion, (ii) a persistence-as-discount theorem making identity continuity a precondition for sustained effort, (iii) an incentive-capability exchange-rate frontier yielding a per-task routing rule, and (iv) a Goodhart bound for multi-channel verification, plus a three-channel decomposition of mechanism gains (allocation, verification, continuation). A 240-run real-LLM experiment with two models and 60 verifiable tasks shows that prompt-level incentive framing has no detectable effect (p = 0.62), whereas runtime allocation raises success on the hardest task family from 0.60 to 1.00 (p = 0.008) at 4.8x price-weighted cost; the exchange-rate theorem predicts the break-even baseline at 0.21. A 20-seed, 3,000-task simulation shows resource ROI rising from 7.09 to 16.67 and gaming falling from 0.333 to 0.115. Code, simulator, and all results are released; live deployment remains future work.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include AI智能体, principal-agent, 提示词工程, arXiv.
Source attribution: curated from arXiv cs.AI (@Jipu Su).
