[目的/意义］AI治理是推进AI赋能社会以及建设负责任人工智能的主要保障手段。文章从人文社会科 学视角研究AI治理，对于AI治理的理论研究和实践应用具有一定的参考价值。[方法/过程］首先，以波士顿咨询、 微软、谷歌、脸书和奔驰集团的AI治理为典型案例进行分析，提出了AI治理的“1+7原则”。其次，采用分层分析 和WSR方法提出了AI治理的实施框架，并将AI治理分为六个基本活动（研发、迭代、运维、问责、训练和共生）、三 个主要层次（技术层、泛技术层和非技术层）和三个不同维度（物理、人理和事理）；接着，探讨由人、人群、智能个 体、智能群体、人机混合个体和人机混合群体组成的新兴社会的人文科学与社会科学属性；最后，提出AI治理亟 待解决的人文社会科学领域的五个研究课题。[结果/结论］虽然AI研发属于自然科学问题，但AI治理应归于人文 社会科学的范畴。AI治理需要重视人文性、社会性、价值理性和交叉学科性。人文社会科学视角下的AI治理需 要重点研究的核心问题有：以人为中心的AI治理体系建设，人类文化与价值观向智能体转移，人机共生的法律、 道德与伦理体系建设，算法经济与智能代理的经济行为以及AI公平性与数字包容。

Abstract: [Purpose/significance] AI governance is the cornerstone for advancing an AI-enabled society and building responsible AI. Humanities and social sciences are the theoretical foundations of AI governance. [Method/process] First, the "1+7 Principles" for AI governance is proposed based on case studies on the AI governance principles of Bos? ton Consulting Group, Microsoft, Google, Facebook, and Mercedes-Benz Group. Secondly, the implementation frame? work of AI governance is proposed using hierarchical analysis and the WSR method. AI governance is divided into six basic activities (R&D, iteration, operation and maintenance, accountability, training, and symbiosis), three main levels (technical technical-related and non-technical layers), and three different dimensions (Wuli, Renli, and Shili dimen? sions); then, the features of AI governance is described via discussing the main roles of AI ecosystem:individuals, groups, intelligent individuals, intelligent groups, human- machine hybrid individuals, and human- machine hybrid groups; Finally, five main topics of humanities and social sciences perspective studies on AI governance are identified. [Result/conclusion] Although the R&D of AI is a natural science issue, AI governance is subjected to the humanities and social sciences. AI governance needs to deal with humanism, sociality, value rationality, and interdisciplinarity studies. The core issues in AI governance from the perspective of humanities and social sciences are: to build a humancentered AI governance system, to transfer human culture and values to intelligent agents, to establish legal, moral, and ethical methods for human-machine co-existing society, algorithmic economy, and agent-based economic behaviors, as well as AI fairness and digital inclusion.