用时间差区分人类、脚本和智能代理，防止AI冒充用户。
aCAPTCHA: Verifying That an Entity Is a Capable Agent via Asymmetric Hardness
- 通过动作、推理、记忆三维度验证智能体能力，利用人与AI处理速度差异
- 设计时间受限的语言理解测试，成功区分三类实体，准确率接近100%
- 无需额外基础设施，可直接嵌入服务作为通用身份验证门禁
随着自主AI代理日益充斥互联网，一个新的安全挑战浮现：‘这个实体是智能代理吗？’ 这是一个新型实体类型验证问题，尚无成熟解决方案。我们基于可验证的智能体能力向量 <x, r, s>（行动、推理、记忆）建立三类实体分类体系（人类、脚本、代理），并利用时间阈值 t 利用人认知与AI处理之间的不对称难度进行区分。通过三个必要性原语定义了智能体能力验证问题（ACVP）。在此基础上，提出 aCAPTCHA（Agent CAPTCHA），一种基于时间约束的安全博弈机制，其安全性依赖于在阈值 t 下 ACVP 的计算难度。我们以时间限制下的自然语言理解为例，实现多轮HTTP验证协议，并通过初步代理测试验证了该协议的正确性与完备性。aCAPTCHA 提供了一种可组合、无需基础设施的准入机制，适用于所有需要实体类型验证的服务场景。
原文摘要 · Abstract (English)
As autonomous AI agents increasingly populate the Internet, a novel security challenge arises: "Is this entity an AI agent?" It is a new entity-type verification problem with no established solution. We formalize the problem through a three-class entity taxonomy (Human, Script, Agent) based on a verifiable agentic capability vector <x, r, s> (action, reasoning, and memory). A timing threshold t exploits the asymmetric hardness between human cognition and AI processing to separate the three classes. We define the Agentic Capability Verification Problem (ACVP) through three necessity primitives, each testing one capability dimension. Building on this foundation, we introduce aCAPTCHA (Agent CAPTCHA), a time-constrained security game for agent admission whose security rests on ACVP hardness under t. We instantiate aCAPTCHA through time-bounded natural-language understanding as a multi-round HTTP verification protocol, and evaluate it with preliminary agent trials that validate the protocol's soundness and completeness. aCAPTCHA provides a composable, infrastructure-free admission gate for any service where entity-type verification is required.
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。