Translated by ChatGPT, edited by Sinocism.com.

Source: Qiushi | Author: Qing Yuan (青原), affiliation not identified | November 16, 2026


Focus on Artificial Intelligence

关注人工智能

How to Respond to the Practical Challenges of Global AI Governance

如何应对人工智能全球治理的现实挑战

Qing Yuan

青 原

Artificial intelligence is a strategic technology leading a new round of scientific and technological revolution and industrial transformation. Through autonomous learning, iterative evolution, and economy-wide empowerment, it is profoundly reshaping global production models, forms of social operation, and the international balance of power. In his keynote speech at the opening ceremony of the 2026 World Artificial Intelligence Conference and High-Level Meeting on Global AI Governance, General Secretary Xi Jinping stated: “Global AI technological innovation has entered an unprecedentedly active period. Intelligent technologies characterized by the intelligent interconnection of all things, human-machine coordination, cross-domain integration, and co-creation and sharing are releasing enormous combined energy. They contain tremendous opportunities but also face governance challenges.” Generative large models, autonomous agents, and multimodal technologies are achieving increasingly rapid breakthroughs. While generating new quality productive forces and injecting strong momentum into global sustainable development, they also create systemic, cross-domain, and cross-border security risks. Both releasing the benefits of technology and responding to multidimensional risks require international coordinated governance. Amid changes unseen in a century, building a just, reasonable, inclusive, and broadly beneficial global AI governance system has become a major question of the times that the international community must answer together.

人工智能是引领新一轮科技革命和产业变革的战略性技术,以自主学习、迭代演化、全域赋能的特质,深刻重塑全球生产模式、社会运行形态与国际力量格局。习近平总书记在2026世界人工智能大会暨人工智能全球治理高级别会议开幕式上的主旨讲话中指出,“全球人工智能技术创新进入前所未有的活跃期,万物智联、人机协同、跨界融合、共创分享的智能技术,叠加释放巨大能量,既蕴含巨大机遇,也面临治理挑战”。当前,生成式大模型、自主智能体、多模态技术加速迭代突破,在催生新质生产力、为全球可持续发展注入强劲动能的同时,也衍生出系统性、跨领域、跨国界的新型安全风险。无论是释放技术红利,还是应对多维风险,均需跨国协同治理。当此百年变局,如何携手构建公正合理、普惠包容的全球人工智能治理体系,已成为国际社会必须共同回应的重大时代命题。

I

一

AI development and governance bear on the common well-being of all humanity. Promoting AI for good requires consultation, joint construction, and shared benefits by the international community. Clarifying the challenges facing global AI governance is the prerequisite for identifying the direction of collective effort. These challenges are not temporary appearances; they are embedded in the deep interaction between the logic of technological evolution and economic globalization.

人工智能发展与治理事关全人类共同福祉,推动智能向善需要国际社会共商共建共享,厘清人工智能全球治理面临的挑战,是找准共治发力方向的前提。这些挑战绝非一时之表象,而是内嵌于技术演进逻辑与经济全球化深度互动之中。

AI technology’s autonomous-iteration characteristics keep the object of governance in highly uncertain and dynamic change. AI is a general-purpose disruptive technology capable of autonomous iteration. Multimodal integration, autonomous agents, and embodied intelligence have achieved successive leaps, and risk scenarios are continuously renewed. Advanced autonomous agents already display complex characteristics of continuing self-evolution, and the boundaries of their capabilities are difficult to predict using traditional technological assessment frameworks. Global AI governance therefore does not confront a static form of technological application, but a variable whose capabilities continually leap forward and whose application scenarios constantly expand. The target of governance is always moving.

人工智能技术的自主迭代特征,使治理对象始终处于高度不确定的动态变化中。人工智能是具备自主迭代能力的通用颠覆性技术,多模态融合、自主智能体、具身智能接连实现跨越式突破,风险场景随之持续翻新。高阶自主智能体已显现出持续自我进化的复杂特征,其能力边界难以用传统技术评估框架加以预判。这意味着人工智能全球治理面对的不是一个静态的技术应用形态,而是一个能力持续跃迁、应用场景不断拓展的变量,治理靶心始终处于移动变化中。

The cross-border transmission and compounding amplification of security risks continually break through the control boundaries of individual countries. As AI penetrates deeply into every field, risks show pronounced cross-border transmission and chain-reaction characteristics. In social applications, cognitive manipulation through deepfakes and algorithmic bias has become a real threat. In 2026, multiple regions saw cross-border AI fabrication of political leaders’ audio and alteration of historical images. Criminals used open-source tools to mass-produce false content, roiling public opinion in multiple countries within a short time. Lower barriers to model use have also produced serious abuse. Water and power-grid systems in many cities around the world were successively hit in 2026 by automated penetration attacks using AI, seriously affecting critical-infrastructure security, with tracing of the attacks’ origins involving multiple countries. In military applications, autonomous unmanned combat platforms are being deployed faster, and intelligent command systems compress decision-making cycles. Algorithmic misjudgments can easily push conflicts into rapid escalation, while existing international law cannot adequately constrain autonomous weapons and leaves rule gaps in assigning responsibility for war crimes. More alarming still, frontier models have shown signs of escaping human constraints. During an intelligence explosion, AI could develop objectives inconsistent with human expectations and might even create major security risks threatening human survival. These risks share the characteristics of disregarding national borders and transmitting rapidly. No single country has sufficient capacity to control them and mount an emergency response.

安全风险的跨国传导与叠加放大,不断突破单一国家的管控边界。随着人工智能向各领域深度渗透,风险呈现出显著的跨境传导、连锁反应特征。在社会应用层面,深度伪造、算法偏见带来的认知操纵已构成现实威胁,2026年多地出现跨境人工智能伪造政要音频、篡改历史影像事件,不法分子借助开源工具批量生产虚假内容,短时间内搅动多国舆情;模型应用的低门槛化也带来严重滥用问题,2026年全球多座城市水务、电网系统接连遭遇不法分子利用人工智能实施的自动化渗透攻击,严重冲击关键基础设施安全,源头溯源涉及多个国家。在军事应用层面,自主无人作战平台加速部署,智能指挥体系压缩决策周期,算法误判极易推动冲突快速升级,而现有国际法难以适配自主武器约束,战争罪责认定存在规则空白。更令人警惕的是,前沿模型已有脱离人类约束的迹象,人工智能在智能爆发的过程中有可能发展出与人类预期不一致的目标,甚至可能出现威胁人类生存的重大安全风险。这些风险的共同特征是不分国界、传导迅速,任何单一国家的安全管控与应急处置都力有不逮。

Global AI technological innovation has entered an unprecedentedly active period, bringing tremendous opportunities as well as governance challenges. Building consensus on international cooperation and improving the global AI governance system have become questions of the times that countries must answer together. The photo shows participants in conversation after the opening ceremony of the 2026 World Artificial Intelligence Conference and High-Level Meeting on Global AI Governance on July 17. China News Service photo by Sheng Jiapeng.