Agentic Analysis for Agentic Infrastructure: An LLM-Powered Pipeline for Comparative Governance of DAO and Corporate AI Protocols
Agentic Analysis for Agentic Infrastructure: An LLM-Powered Pipeline for Comparative Governance of DAO and Corporate AI Protocols
代理基础设施的代理分析：一种用于 DAO 与企业 AI 协议比较治理的 LLM 驱动流水线
Abstract: As AI agent protocols proliferate, the governance structures shaping their interoperability standards remain empirically underexamined. We introduce an LLM-powered comparative pipeline for large-scale governance discourse analysis, integrating automated annotation, neural topic modeling, and multi-layer network analysis to study socio-technical power structures at scale.
摘要： 随着 AI 代理协议的激增，塑造其互操作性标准的治理结构在实证研究方面仍未得到充分探讨。我们引入了一种由大语言模型（LLM）驱动的比较流水线，用于大规模治理话语分析，通过整合自动标注、神经主题建模和多层网络分析，在大规模尺度上研究社会技术权力结构。
We validate it on two contrasting standards for agent interoperability: ERC-8004 (permissionless, on-chain) and Google A2A (corporate-led). Analyzing 4,323 governance participation records, we combine LLM-assisted coding, topic modeling, and multi-layer network analysis to examine how institutional design shapes thematic priorities and community structure.
我们在两种截然不同的代理互操作性标准上验证了该方法：ERC-8004（无需许可、链上）和 Google A2A（企业主导）。通过分析 4,323 条治理参与记录，我们结合了 LLM 辅助编码、主题建模和多层网络分析，以考察制度设计如何塑造主题优先级和社区结构。
We find that while governance form influences substantive focus, both regimes exhibit comparable levels of participation inequality and community fragmentation. Discourse alignment is denser in the permissionless setting, suggesting that open governance may foster greater thematic convergence despite decentralized participation.
研究发现，尽管治理形式会影响实质性关注点，但两种机制在参与不平等和社区碎片化方面表现出相当的水平。在无需许可的环境中，话语的一致性更为紧密，这表明开放式治理尽管存在去中心化的参与，却可能促进更强的主题趋同。
These findings illustrate how LLM-assisted methods can advance the empirical study of technology governance, with implications for designing more equitable agentic AI standards. All data and code are openly available.
这些发现阐明了 LLM 辅助方法如何推动技术治理的实证研究，并为设计更公平的代理 AI 标准提供了启示。所有数据和代码均已公开。