Abstract: The logic of security confrontation in international politics and the competition of interests among states have created significant spillover effects on global artificial intelligence (AI) governance, leading to a pronounced panpoliticization tendency in technological governance issues. The competitive domains of AI governance have extended beyond the formulation of rules to encompass areas such as the development of computational power, data security, and algorithm performance. At the same time, discursive struggles surrounding the values of governance continue to affect the normative foundations of global governance, exacerbating the fragmentation and antagonism within the governance system. To mitigate the potential crises arising from this tendency toward panpoliticization, global AI governance needs to return to the original technocentric purpose. This necessitates a thorough analysis of the impacts of deep developments in humanmachine interaction, the highlevel aggregation of information, and the evolution of anthropomorphic characteristics of intelligence on governance. Relevant actors should transcend existing divisions of interest, discard the logic of competition and confrontation, and place the healthy development of technology at the core starting point of governance discussions. This will promote the depoliticization of global AI governance, aiming to establish a more inclusive, collaborative, and efficient governance system.

Key words: artificial intelligence, global governance, panpoliticization, governance competition

摘要： 国际政治的对抗逻辑及国家间的利益竞争对全球人工智能治理产生了显著的外溢效应，致使技术治理议题呈现泛政治化倾向。当前，人工智能治理的竞争领域不再局限于规则层面的较量，进一步扩展至算力发展、数据安全、算法性能等多个方面。同时，围绕治理价值的话语博弈正持续冲击着全球治理的价值基础，加剧治理体系的分裂与对垒。为缓解其泛政治化倾向带来的潜在危机，全球人工智能治理亟须回归技术中心的治理初衷，剖析人机交互的深度发展、信息的高度聚合以及智能拟人特征的演进对治理产生的影响。相关行为体应突破既有利益分隔，摒弃竞争对抗逻辑，将技术的健康发展作为治理讨论的核心出发点，推动全球人工智能治理的去政治化，以期构建更加包容、协同、高效的治理体系。

关键词: 人工智能, 全球治理, 泛政治化, 治理竞争