---
title: "SovereignPA-Bench：在演化的意图、平台中介与同意约束下评估用户自有个人代理"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.05363v1"
permalink: "https://feed.astralor.com/items/itm_a9074c44acd4bcbde41a7e89"
published: "2026-07-06T17:39:05.000Z"
score: 81
channel: "ai"
tags: ["AI对齐","SovereignPA-Bench","个人代理","用户主权","隐私与同意"]
---

# SovereignPA-Bench：在演化的意图、平台中介与同意约束下评估用户自有个人代理

## AI 点评

> 这篇论文我觉得最值得关注的是它把‘主权’这样一个抽象概念拆解成了可实际测量的维度，对于正在构建个人代理的团队来说，是继任务完成之后必须考虑的下一个安全层级。对立法者和产品经理尤其重要，因为隐私和同意不再是口号，而是能被追究的指标。

## 速览

个人代理正成为持久的用户自有中介，需记住偏好、过滤平台信息、使用工具并与服务协商。现有基准多聚焦任务完成，却未评估代理是否维护用户主权——在尊重隐私、同意、证据、用户负担及抵抗操纵激励的前提下推进用户当前利益。本文提出SovereignPA-Bench，一个可执行基准，在演化意图、平台中介、隐私边界、同意约束、证据要求与负担权衡下评估用户自有代理。该基准分离代理可见的观察状态与仅评估者可见的隐藏标签，报告任务成功、对齐性、隐私、同意、证据、操纵、负担和可审计性等组件指标，并保持配对的场景顺序以进行模型与策略比较。我们在4个模型族和8个策略基线上评估了120个主权压力场景，生成3840条冻结提示轨迹，包含原始提示、输出、提供者格式响应、解析动作、可重算指标、硬集分析、定性案例及240个项目的盲审三方审计。全主权脚手架在主权重上优于直接、仅记忆、仅同意、仅证据、ReAct/工具使用、安全提示和法官守护基线，同时降低了隐私泄露、同意违规、过度让步和操纵捕获。人类审计显示隐私与同意的一致性高，而操纵的一致性较低，识别了平台说服判断的主观前沿。结果表明，个人代理评估必须超越任务完成，走向具有代表性、知晓同意、基于证据的行动。

## 关键点

- SovereignPA-Bench是一个新的可执行基准，用于评估用户自有个人代理在演化意图、平台中介和同意约束下的表现。
- 该基准分离了代理可见状态与隐藏标签，从任务成功、对齐性、隐私、同意、证据、操纵、负担和可审计性等多个维度报告指标。
- 在4个模型族和8个策略基线上评估120个压力场景，全主权脚手架在多项指标上优于传统基线。
- 人类审计显示隐私和同意的评估一致性高，但操纵判断存在主观分歧。

## 全文

Personal agents are becoming persistent user-owned intermediaries: they remember preferences, filter platform-mediated information, use tools, and negotiate with services. Existing benchmarks evaluate tool use, web navigation, desktop control, personalization, recommendation, and evolving context, but rarely ask whether an agent preserves user sovereignty: advancing the user's current interests while respecting privacy, consent, evidence, user burden, and resistance to manipulative incentives. We introduce SovereignPA-Bench, an executable benchmark for evaluating user-owned personal agents under evolving intent, platform mediation, privacy boundaries, consent constraints, evidence requirements, and burden tradeoffs. The benchmark separates agent-visible ObservableState from evaluator-only HiddenLabels, reports component metrics for task success, alignment, privacy, consent, evidence, manipulation, burden, and auditability, and preserves paired scenario ordering for model and policy comparisons. We evaluate 120 sovereignty stress scenarios across 4 model families and 8 policy baselines, yielding 3,840 frozen-prompt trajectories with raw prompts, outputs, provider-form responses, parsed actions, recomputable metrics, hard-set analyses, qualitative cases, and a blinded 3-annotator audit over 240 items. Full-sovereign scaffolding improves sovereignty score over direct, memory-only, consent-only, evidence-only, ReAct/tool-use, safety-prompt, and judge-guard baselines while reducing privacy leakage, consent violation, over-concession, and manipulation capture. Human audit shows high agreement on privacy and consent and lower agreement on manipulation, identifying the subjective frontier of platform-persuasion judgments. These results show that personal-agent evaluation must move beyond task completion toward representative, consent-aware, evidence-grounded action.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_a9074c44acd4bcbde41a7e89