Three of China's Leading Scholars Discuss the Future of Global AI Governance

Also in this edition: United Front Works Department on Significance of Ethnic Unity & Progress Law

Hi folks,

In today’s edition, I am breaking down two long pieces from the paper. The first is an interview with experts on AI governance. The second is an article on the ethnic unity and progress law.

On the theory page today, there’s a conversation on global governance of artificial intelligence (AI), with three scholars. These are Xue Lan, Dean of the Institute for AI International Governance at Tsinghua University, Tang Shiqi, Dean of the School of International Studies at Peking University, and Song Guoyou, Professor at the Institute of International Studies at Fudan University.

I found this to be an extremely interesting conversation. There are some broad takeaways that I can highlight.

• First, all of them seem to think that there is very little room at present for some sort of grand bargain among great powers on AI governance. This is a product of the stage at which technological development is at, but also because of the nature of great power competition, which has eroded trust, and the role of the private sector across countries. Tang proposes finding some sort of a balance between national security, economic competition, and openness and mutual trust. But this is abstract. That said, some areas for cooperation that they all propose are jointly managing and controlling military applications of AI, developing AI for healthcare and for mitigating global disasters, platforms for addressing climate change, global infectious disease early warning models, and crop health management systems, etc.

• Second, there’s a general unease that is expressed throughout the interview when it comes to the power of the private sector, and its implications for political power and on governance. For instance, Tang cautions that enterprises command control over “foundational models, control massive amounts of user data, and wield powerful influence”. Song warns that private enterprises are seeking “market-leading advantages and excess returns” and harbour concerns related to “government-led joint construction of an AI global governance system.”

• Finally, in terms of global AI governance, they do propose some interesting principles. But, there is deep scepticism about translating some of these ideas into effective governance outcomes. This is because of the fractured nature of the institutional mechanisms of governance, but also capability disparities among countries, different needs and interpretations and inability to arrive at some common understandings. What is noteworthy is that they all back a UN-centered framework.

Anyway, there’s a lot more that one can distil, but you can do that by going through the full text below.

The first question is about the particular challenges related to AI governance in terms of arriving at some sort of an arrangement globally.

Xue says: “Promoting the improvement of global governance of artificial intelligence is an important component of practicing GGI. The characteristics of global governance of AI stem from the rapid iterative development of AI technology, while its impacts are extremely broad and far-reaching, and the international community has not made sufficient preparations in thinking and action. This objective situation leads to some characteristics in AI global governance compared with relatively mature global governance issues. For example, issues such as addressing climate change challenges reached consensus at a time when the atmosphere for international cooperation was relatively favourable. In contrast, the explosive development of AI technology coincides with the continuous rise of geopolitical risks, and from its very inception, global governance of AI has faced a complex situation of great-power games and technological competition. A few countries have engaged in ‘decoupling and breaking chains’, erecting ‘small yard, high fences’ in areas such as technological development, talent cultivation, and data interconnection, attempting to monopolize competitive advantages through technological blockades, thereby damaging the previously closely connected global scientific research and industrial collaboration networks. Against this backdrop, the core actors of AI global governance, composed of technologically leading countries, are not ‘cooperators’ but ‘competitors’, which has weakened the trust foundation for AI global governance cooperation, resulting in the international community lacking a cooperative foundation and collaborative mechanisms when responding to the risks and challenges brought by AI technological development. Unilateralism and protectionism have given rise to a landscape where technological barriers and rules barriers coexist, leading to the fragmentation of the AI global governance system, weakening the universal binding force of rules, and causing AI global governance to fall into the predicament of ‘divide and rule’. 推动完善人工智能全球治理，是践行全球治理倡议的重要内容。人工智能全球治理的特点，源于人工智能技术发展快速迭代，而其影响又极为广泛深远，国际社会没有在思想和行动上做好充分的应对准备。这一客观情况，导致与较为成熟的全球治理议题相比，人工智能全球治理有一些特点。比如，应对气候变化挑战等议题是在国际合作氛围较好的时候达成了共识。与此不同，人工智能技术的爆发式发展恰与地缘政治风险持续上升同步，人工智能全球治理从诞生之初便面临大国博弈、技术竞争的复杂局面。个别国家大搞“脱钩断链”，在技术开发、人才培养、数据互联等方面筑起“小院高墙”，试图以科技封锁垄断竞争优势，破坏了原本联系较为紧密的全球科研及产业协作网络。在这种背景下，由技术领先国家构成的人工智能全球治理核心主体之间不是“合作者”，而是“竞争者”，削弱了人工智能全球治理合作的信任基础，导致国际社会在应对人工智能技术发展所带来的风险和挑战时，缺乏合作根基与协作机制。单边主义、保护主义催生出技术壁垒与规则壁垒并存的格局，导致人工智能全球治理体系碎片化，削弱了规则的普遍约束力，导致人工智能全球治理陷入“分而治之”的困境.

For Tang, the nature of AI governance is defined by two aspects.

First, AI is not merely an object to be governed by decisions, but partially participates in decision-making itself. Because of its outstanding information processing and handling capabilities, AI has already become a means of assisting decision-making in many fields. In the future, decision-makers’ dependence on AI will generally continue to increase. The authenticity and effectiveness of the information provided by AI will have an increasingly large impact on the outcomes and quality of decisions. However, the AI technology used to assist decision-making is typically not in the hands of decision-makers, but is controlled by enterprises within the industry that possess significant advantages and powerful resource-integration capabilities. They dominate foundational models, control massive amounts of user data, and wield powerful influence. Therefore, how to ensure that AI achieves comprehensiveness and objectivity in the information supply link is increasingly becoming an important issue. 　　一是人工智能不仅是被决策的对象，而且部分地参与了决策。人工智能因其出色的信息加工和处理能力，已经在诸多领域成为辅助决策的手段。未来，决策者对人工智能的依赖总体上会持续增加。人工智能提供信息的真实性和有效性对决策的结果和质量会产生越来越大的影响。然而，用于辅助决策的人工智能技术通常并不掌握在决策者手中，而是由行业内具备显著优势和强大资源整合能力的企业所掌控，它们主导基础模型、掌握海量用户数据，具有强大的影响力。因此，如何确保人工智能在信息供应环节做到全面、客观，日益成为一个重要问题.

Second, the objects of governance possess fluidity and virtuality. AI is not like trade exchanges or climate change, which can manifest their characteristic changes and development trends through some observable, controllable entities. The objects involved in AI global governance are computing power, algorithms, data, and models; they themselves are technology, not the results produced by technology, and therefore, AI lacks a clear ‘governance anchor point’. In addition, AI exists and flows freely in cyberspace, to a large extent detached from people’s physical grasp and control. The information and data collected by AI, as well as the specific ways of processing such information and data, are not entirely within human controllable scope. Generative AI possesses a closed-loop mechanism of ‘learning—optimisation—regeneration’, and its behaviour may exceed the developers’ initial design, leading to a disconnect between intention and outcome. 二是治理对象具有流动性和虚拟性。人工智能并不像贸易交往、气候变化那样，能够通过一些可观察、可控制的实体来展现其特征变化与发展趋势。人工智能全球治理涉及的对象是算力、算法、数据和模型，它们本身是技术，而非技术产生的结果，因此人工智能缺乏清晰的“治理锚点”。此外，人工智能存在并自由流动于网络空间，在很大程度上游离于人们物理性的把握和控制之外。人工智能所采集到的信息和数据，以及对这些信息和数据的具体处理方式并不完全在人的可控范围。生成式人工智能具备“学习—优化—再生成”的闭环机制，其行为可能超出开发者的初始设计，导致意图与结果脱节.

Song contends that “compared with relatively mature global governance issues, the characteristics of global AI governance have at least three aspects. The first is the imbalance of governance impact. Climate change is global, and all countries are greatly affected. However, the development and impact of AI are not balanced. Some countries have already felt the intense impact brought by AI, while some countries may believe they are unaffected or less affected. The second is the unpredictability of governance pathways. Since AI itself is still in an early stage of development and has not yet taken shape, specific matters such as governance priorities and rules are difficult to implement at this stage, and governance pathways are not clear. The third is the high sensitivity of technological competition. As an important driving force of a new round of scientific and technological revolution and industrial transformation, AI is a strategic commanding height of great-power competition. The high sensitivity of technological competition leads some countries to have strong competitive and confrontational thinking, and insufficient awareness of cooperation and mutual benefit. 宋国友：与相对成熟的全球治理议题相比，人工智能全球治理的特点至少有如下3个方面：一是治理影响的不均衡性。气候变化是全球性的，所有国家都受到巨大影响。然而，人工智能的发展和影响并不均衡。一些国家已经感受到人工智能带来的剧烈冲击，而部分国家可能认为自身不受影响或影响较小。二是治理路径的非预知性。由于人工智能本身还处于早期发展阶段，尚未定型，对其治理的重点和规则等具体事项现阶段难以落实，治理路径并不清晰。三是技术竞争的高敏感性。人工智能作为新一轮科技革命和产业变革的重要驱动力量，是大国竞争的战略制高点。技术竞争的高敏感性导致个别国家竞争和对抗思维强烈，合作和互利意识不足.

The second question is about China’s Global AI Governance Initiative, and the challenges that the joint construction of a global AI governance system faces.

Xue makes three points:

• First, the international community lacks consensus on prominent issues in global AI governance. For example, how should the potential risks of AI be understood? What red lines need to be drawn in AI development? How should governments achieve a dynamic balance between promoting innovation and preventing risks? There are still large differences among parties, posing challenges for the international community in implementing the Global AI Governance Initiative. At the same time, how to bridge the increasingly widening intelligence divide is also an important issue that China and other developing countries hope will attract the attention of the international community.

• Second, AI technology is in a stage of explosive development, and its technological breakthroughs often precede the formulation of governance rules. This ‘time lag’ causes governance to always be ‘chasing’ technology, forming a governance difficulty. For example, in the face of risks brought by new technologies such as generative AI and embodied intelligence, current governance rules show gaps or lag. Therefore, AI safety governance methods such as agile governance advocated by China have received high attention from the international community.

• Third, at the global level, there appear to be many governance mechanisms, such as the UN, the OECD, and others are promoting the formulation of relevant guidelines, and countries have their own standards and norms. However, these mechanisms lack overall coordination, and there are overlaps and even conflicts among them, forming a ‘complex of mechanisms’, leading to complicated rules that are difficult to implement. Countries face a ‘maze of rules’, execution costs are high, and actual effectiveness is greatly reduced.

薛澜：第一，国际社会在人工智能全球治理的突出问题上缺乏共识。比如，如何认识人工智能的潜在风险？需要在人工智能发展中划定哪些红线？政府如何在促进创新与防控风险之间取得动态平衡？各方仍存在较大分歧，给国际社会落实《全球人工智能治理倡议》带来挑战。同时，如何弥合日益加剧的智能鸿沟也是中国和其他发展中国家希望能够引起国际社会关注的重要问题。第二，人工智能技术正处于爆发式发展阶段，其技术突破往往领先于治理规则的制定，这种“时间差”导致治理始终在“追赶”技术，形成治理难题。比如，面对生成式人工智能、具身智能等新技术带来的风险，当前治理规则存在缺位或滞后的现象。因此，中国所倡导的敏捷治理等人工智能安全治理方式得到国际社会高度关注。第三，全球层面看似治理机制众多——联合国、经济合作与发展组织等都在推动制定相关准则，各国还有各自的标准和规范。但这些机制缺乏统筹协调，彼此间存在重叠甚至冲突，形成了一个“机制复合体”，导致规则繁杂且难以落地，各国面临“规则迷宫”，执行成本高企，实际效能大打折扣.

Tang argues that at present, technological nationalism is on the rise around the world. Some countries are prioritising their own technological security and industrial chain resilience rather than global public interests. At the enterprise level, AI development is generally in a competitive state lacking order and rules, and is intertwined with competition among countries, bringing great difficulty to establishing a global AI governance system. In terms of governance mechanisms, different countries also have a series of divergences on key issues such as cross-border data flows, content review, and government regulatory authority. Reducing contradictions and conflicts requires various international actors, especially major powers, to enhance mutual trust and cooperation, establish a relatively complete AI monitoring and early warning mechanism, as well as mechanisms for sharing interests and responsibilities, and on this basis enhance consensus as much as possible. 唐士其：目前，世界范围内的技术民族主义抬头，一些国家优先考虑的是本国的技术安全与产业链韧性，而非全球公共利益。从企业层面看，人工智能的开发总体上处于一种缺乏秩序和规则的竞争状态，并且与各个国家之间的竞争交织在一起，为建立人工智能全球治理体系带来很大困难。在人工智能治理机制方面，不同国家在数据跨境、内容审核、政府监管权限等关键问题上也存在一系列分歧。减少矛盾和冲突，需要各种国际行为体，特别是大国之间增进互信与合作，建立一套相对完整的人工智能监测、预警机制，以及利益和责任的分担机制，在此基础上尽可能增强共识.

Song says that “analysing from the perspective of the co-building actors, there are three difficulties. First, some countries pursue unilateralism and protectionism, hindering the joint construction of a global AI governance system. Second, some countries with less developed AI technology are not highly enthusiastic about co-building. These countries have not truly experienced the impact brought by AI, lack understanding of its opportunities and risks, and lack a strong willingness to participate in global AI governance. Third, the private sector participating in the joint construction of the AI global governance system harbors concerns. Large AI enterprises and other private-sector entities are both important actors in global AI governance and simultaneously objects being governed. The relevant private sector hopes to center on enterprise development, with profit as its goal, in order to obtain market-leading advantages and excess returns, and still has doubts about and resistance toward the government-led joint construction of an AI global governance system. 宋国友：从共建主体的角度加以分析，困难有3点。一是个别国家奉行单边主义、保护主义，阻碍了人工智能全球治理体系共建。二是部分人工智能技术欠发达的国家对于共建热情不高。这些国家还没有真正经历人工智能带来的冲击，对其机遇和风险缺乏认识，缺少参与人工智能全球治理的迫切意愿。三是参与人工智能全球治理体系共建的私营部门存在担忧。大型人工智能企业等私营部门既是全球人工智能治理的重要主体，同时也是被治理的客体。相关私营部门希望以企业发展为中心，以利润为目标，以便获得市场领先优势和超额回报，对于政府主导的共建人工智能全球治理体系仍有疑虑和抵触.

The third question basically is about what kind of governance concepts should be promoted globally to achieve coordinated advancement of technological development and governance effectiveness?

Xue makes four points:

• First, it is necessary to adhere to a people-centered development philosophy. Ultimately, the development of AI must serve human society. We have consistently emphasised ‘technology for good’, which means taking human well-being as the fundamental goal of technological development. Whether in technology design or rule-making, everything must revolve around human needs, ensuring that technological development does not deviate from the trajectory of human social development, helping to bridge the North–South development gap, and promoting all countries to share the achievements of AI development.

• Second, establish a governance foundation of equal dialogue. Global governance of AI cannot be decided by a few technologically powerful countries. Countries differ in capabilities and size, but they should participate equally in the formulation of governance rules and adhere to the global governance concept of joint consultation, joint construction, and joint sharing. We need to abandon unilateralism, build multilateral platforms, fully respect the voices of different countries, especially the demands of developing countries, and enhance their representation and voice. Only through inclusiveness and mutual learning can a governance framework with real consensus be formed.

• Third, it is necessary to focus on an action orientation and actively promote the establishment of a governance pathway for inclusive development. At present, there are clear technological gaps among countries. If these gaps are allowed to widen, the governance divide will only deepen. Therefore, China particularly emphasises strengthening capacity building, adhering to systematic planning and overall advancement, and through means such as technical assistance and knowledge sharing, promoting countries of the Global South to share the dividends of AI development.

• Finally, construct a risk prevention and control system of shared security governance. AI risks are global; a technological vulnerability in one country may trigger chain reactions worldwide. Therefore, AI safety governance cannot be ‘each sweeping snow from their own doorstep’. AI safety should be provided as a global public good, with countries jointly developing safety technologies, formulating unified standards, and establishing emergency coordination mechanisms.

薛澜：首先，必须坚持以人为本的发展理念。人工智能发展归根结底要服务人类社会。我们一直强调“技术向善”，就是要把人的福祉作为科技发展的根本目标。无论是技术设计还是规则制定，都要围绕人的需求展开，确保技术发展不偏离人类社会发展轨道，助力弥合南北发展鸿沟，推动各国共享人工智能发展成果。其次，确立平等对话的治理基础。人工智能全球治理不能由少数技术强国说了算。各国能力有强弱，体量有大小，但在治理规则制定上应该平等参与，坚持共商共建共享的全球治理观。我们需要摒弃单边主义，搭建多边平台，充分尊重不同国家的声音，特别是发展中国家的诉求，提高发展中国家的代表权和发言权。只有包容互鉴，才能真正形成有共识的治理框架。再者，必须注重行动导向，积极推动确立普惠发展的治理路径。现在各国技术差距明显。如果放任这种差距扩大，治理鸿沟只会更深。所以，中国特别强调要加强能力建设，坚持系统谋划、整体推进，通过技术援助、知识共享等方式，推动全球南方国家共享人工智能发展红利。最后，构建安全共治的风险防控体系。人工智能风险是全球性的，一个国家出现技术漏洞可能引发全球连锁反应。因此，人工智能安全治理不能“各扫门前雪”。要把人工智能安全作为全球公共产品来提供，各国共同研发安全技术、制定统一标准、建立应急协作机制.

Tang outlines three principles:

• First, upholding human-centeredness. Traditionally, we have believed that what makes humans human is to a great extent that humans possess rationality, and the appearance of an AI entity that in certain respects far surpasses human wisdom, especially logical thinking ability, has without doubt shaken humanity’s traditional standing. Faced with AI, in what aspects we should safeguard human dignity, stimulate human creativity, and at the same time maintain basic humanistic bottom lines is a fundamental issue that all humanity needs to face collectively, discuss together, and form basic consensus and value norms. In any case, AI can only be a tool of humans; its use must be people-centered. This is an important principle.

• Second, upholding mutually beneficial cooperation. At present, the development and use of AI exhibit characteristics of multi-centeredness and decentralization, while also showing clear global imbalance. Therefore, on the one hand, it is necessary to complement each other’s advantages and avoid redundant construction and waste; on the other hand, it is necessary to enable countries and regions relatively lagging in AI development to also enjoy the great convenience brought by AI and narrow the digital divide. Both aspects require advocating the spirit of mutually beneficial cooperation in the process of global AI governance.

• Third, upholding openness and mutual trust. AI technology is built on the basis of large language models; the more open information is and the more freely it flows, the faster AI will develop. At the same time, AI’s application in fields such as the military and national security, as well as its enormous economic potential, make it a key domain of great power competition. Since this situation is difficult to fundamentally change in the short term, on the issue of global AI governance, efforts should be made to find a balance point, i.e., a balance between national security, economic competition, and openness and mutual trust.

唐士其：在这个问题上，我觉得必须坚持3点。一是坚持以人为本。传统上我们认为人之为人，很大程度上是因为人拥有理性，而一种在某些方面远远超越了人的智慧，特别是逻辑思维能力的人工智能体的出现，毫无疑问撼动了人类传统的地位。那么面对人工智能，我们究竟应该在什么方面维护人类的尊严、激发人类的创造力，同时保持基本的人文底线，就是一个需要全人类集体面对、共同探讨，并形成基本共识与价值规范的根本性问题。无论如何，人工智能只能是人类的工具，对人工智能的利用必须以人为本，这是一个重要原则。二是坚持互利合作。目前，人工智能的开发和利用都体现出多中心、分散化的特点，同时在世界范围呈现出明显的不均衡性。因此，一方面，要优势互补、避免重复建设和浪费；另一方面，要让人工智能研发相对落后的国家和地区也能够享受人工智能带来的巨大便利、缩小数字鸿沟。这两方面都需要在人工智能全球治理的过程中提倡互利合作的精神。三是坚持开放互信。人工智能技术建立在大语言模型的基础上，信息越开放，越能够自由流动，人工智能的发展就会越快。人工智能在军事、国家安全等领域的运用，及其巨大的经济潜能，又使人工智能成为大国博弈的关键领域。由于短期内这种局面难以得到根本性改变，所以在人工智能全球治理的问题上，应该努力寻找一个平衡点，即国家安全、经济竞争与开放互信的平衡点.

Song says that “AI technology is developing with each passing day. Global AI governance should ensure that no matter how AI technology iterates, it conforms to the common interests of all humanity and that AI development serves the well-being of all humankind. Specifically, governance concepts such as open intelligence, inclusive intelligence, universally beneficial intelligence, and safe intelligence should be strengthened. Open intelligence means that global AI governance should prevent a few major powers from forming blocs around AI and imposing blockades on other countries. Inclusive intelligence means that global AI governance should incorporate the civilizations of all humanity, and should not inject civilisational exclusivism into AI. Universally beneficial intelligence means enabling AI to benefit all humanity, allowing all countries in the world to gain from AI technological innovation. Safe intelligence means actively preventing various risks and challenges that AI may pose to national security, ensuring that it does not threaten humanity. Only by running these four major concepts through the entire process of global AI governance and making them firm ‘guardrails’ of governance can coordinated advancement of technological development and governance effectiveness be achieved. 　宋国友：人工智能技术发展日新月异，人工智能全球治理应确保人工智能技术无论如何迭代，都能符合全人类的共同利益，让人工智能发展服务于全人类福祉。具体而言，应强化开放智能、包容智能、普惠智能和安全智能等治理理念。开放智能，是指人工智能全球治理要防止个别大国将人工智能阵营化，对其他国家搞封锁。包容智能，是指人工智能全球治理应该把全人类的文明都包含在内，而不能在人工智能中注入文明排他主义基因。普惠智能，是指要让人工智能造福于全人类，让世界各国能够在人工智能技术创新中获益。安全智能，是指要积极防范人工智能可能引发的各种危害国家安全的风险和挑战，确保其不威胁人类。只有将这四大理念贯穿于人工智能全球治理全过程，成为牢固的治理“护栏”，才能实现技术发展和治理效能的协同推进.

The next question is about how should countries break the gridlock on AI governance to achieve the Chinese proposition of “promoting the construction of an inclusive and fair multilateral global digital governance system”?

We only have responses from Xue and Song.

Xue says that first, it is necessary to firmly support the United Nations in playing its role as the main channel, and support the establishment under the UN framework of an ‘independent international scientific panel on AI’ and a ‘global dialogue on AI governance’ mechanism. At the same time, it is essential to actively encourage and participate in various other bilateral and multilateral dialogue mechanisms, and explore a path of open consultation, diverse participation, mutual trust, and co-governance. Second, while encouraging technological innovation and industrial application, establish through international cooperation an AI risk testing and assessment system, improve data protection norms, and seek a balance between development and security. In addition, we must actively advocate open cooperation, help countries around the world, especially countries of the Global South, strengthen capacity building, and promote the shared enjoyment by all countries of the fruits of AI development. 薛澜：首先，必须坚定支持联合国发挥主渠道作用，支持在联合国框架下建立“人工智能独立国际科学小组”和“人工智能治理全球对话”机制。同时，积极鼓励和参与其他各种双边和多边对话机制，走出一条开放协商、多元参与的互信共治之路。其次，在鼓励技术创新与产业应用的同时，通过国际合作建立人工智能风险测试评估体系、完善数据保护规范，寻求发展与安全的平衡。此外，我们要积极倡导开放合作，帮助世界各国特别是全球南方国家加强能力建设，推动人工智能发展成果由各国共享.

Song says that ‘international cooperation on AI can break through geopolitical barriers from the following aspects.

• First, cooperation on key issues. Cooperation can be carried out first around major issues that pose the greatest risks to humanity and have the strongest public goods attributes; for example, jointly managing and controlling military applications of AI, and developing AI for healthcare and for mitigating global disasters.

• Second, carry out pragmatic cooperation under existing multilateral mechanisms to enhance confidence in cooperation. For example, supporting the UN-initiated ‘global dialogue on AI governance’ mechanism and strengthening rule-based governance of AI within the UN framework.

• Third, promote bilateral and small multilateral cooperation as useful supplements to multilateral cooperation. Such cooperation is more targeted and flexible, and can address AI-related issues of common concern among members, thereby promoting positive outcomes in AI governance and gradually expanding consensus.

• Fourth, encourage market cooperation in the private sector. Market cooperation is an important component of international cooperation. Under the premise of complying with relevant regulatory rules, private sector actors should be encouraged to carry out various forms of commercial cooperation, consolidate the foundation of cooperation at the market level, build models of cooperation, break geopolitical barriers, and promote governance cooperation among countries from the bottom up.

宋国友：习近平总书记强调：“进一步拓展人工智能合作，加强信息交流和技术合作，共同做好风险防范，形成具有广泛共识的人工智能治理框架和标准规范，不断提升人工智能技术的安全性、可靠性、可控性、公平性。”开展人工智能国际合作可从以下几方面突破地缘政治壁垒。一是重要议题合作。可围绕人工智能对人类最具风险和最具公共产品属性的重要议题先行开展合作。前者如合作管控军事应用人工智能等，后者如发展医疗健康人工智能和减轻全球灾害人工智能等。二是开展现有多边机制下的务实合作，增强合作信心。如支持联合国启动的“人工智能治理全球对话”机制，在联合国框架内加强人工智能规则治理，推动相关国际合作。三是推动双边、小多边等合作，作为人工智能国际多边合作的有益补充。双边和小多边合作更具针对性和灵活性，可在双边和小多边成员共同关心的人工智能相关问题方面开展合作，推动人工智能治理取得积极成果，逐渐扩大合作共识。四是鼓励私营部门的市场合作。市场合作是国际合作的重要组成部分，应积极鼓励私营部门在符合相关监管规则等前提下开展各种形式的商业合作，在市场端夯实合作基础、构建合作样板，打破地缘政治藩篱，自下而上地推动国家间治理合作.

The final question is about how can one ensure equal participation of Global South countries in AI governance?

Xue says that “the root of the dual challenges faced by Global South countries lies in weak education and research foundations, lack of talent support, insufficient scientific and technological capabilities, and backward information infrastructure. Under such circumstances, technological empowerment is out of the question. Therefore, to avoid the continued widening of the AI technology divide, high attention must be paid to the education and scientific and technological capacity building of Global South countries, attention must be paid to their information infrastructure construction, and emphasis must be placed on ‘teaching people to fish’. At the level of participation in AI global governance decision-making, it is necessary to strengthen the cultivation of governance capacity in Global South countries, which complements the practice of AI technological application. Inclusive development of AI can both enable Global South countries to enjoy the dividends of AI benefiting humanity and enhance their capacity to participate in global AI governance. 薛澜：全球南方国家面临双重挑战的根源在于教育与研究基础薄弱，缺乏人才支撑，科技能力不足，信息基础设施落后。在这样的情况下技术赋能无从谈起。因此，为避免人工智能技术鸿沟继续扩大，要高度关注全球南方国家的教育及科技能力建设，关注其信息基础设施建设，重视“授人以渔”。在参与人工智能全球治理决策层面，需要加强全球南方国家治理能力的培养，这与人工智能技术应用的实践相辅相成。人工智能普惠发展既可以使全球南方国家享受人工智能造福人类的红利，也可以增强其参与人工智能全球治理的能力.

Tang says that equal participation of Global South countries in global AI governance can be ensured and promoted from three levels.

• First is the technological level. On development issues, major powers have more space for cooperation; for example, jointly building global public-goods-type AI systems, including carbon emissions monitoring platforms for addressing climate change, global infectious disease early warning models, and crop health management systems, and opening them to Global South countries. At the same time, help Global South countries build their own AI systems through various means, cultivate relevant talent, and form their own development capabilities.

• Second is the rules level. In the process of formulating global AI governance rules, fully incorporate participation from Global South countries and ensure fair allocation of representation quotas, so that their demands can influence the basic direction of future AI development.

• Third is the value level. On the one hand, AI powers should avoid exporting AI systems that embed particular value systems to Global South countries through cloud services, pre-trained models, etc., thereby creating new forms of dependency; on the other hand, future AI development should fully reflect the cultural and value factors of Global South countries. Since current AI development is based on human language systems, strictly speaking it is not completely neutral at the level of values. The characteristics of civilisation are reflected not only in values and ethics but also in ways of thinking, all of which are embedded in a nation’s language system. An important reservoir for future AI development lies in the linguistic and cultural resources of non-Western countries. If these resources can be incorporated into AI corpora, this will greatly enhance AI’s level of intelligence and also provide new possibilities for the development of Global South countries and human civilization.

唐士其：习近平总书记指出：“要加强人工智能国际治理和合作，确保人工智能向善、造福全人类，避免其成为‘富国和富人的游戏’。”可以从3个层面保障和促进全球南方国家平等参与人工智能全球治理进程，享受人工智能发展带来的成果。第一是技术层面。在发展问题上，大国有更多合作的空间，比如可以共同建设全球公共产品型人工智能系统，包括应对气候变化的碳排放监测平台、全球传染病早期预警模型、农作物健康管理系统等，并向全球南方国家开放。同时，通过各种方式帮助全球南方国家建立自己的人工智能系统，培养相关人才，并形成自身的发展能力。第二是规则层面。在人工智能全球治理的规则制定过程中充分吸收全球南方国家参与，确保代表名额分配公平，以保证全球南方国家的诉求能够对未来人工智能发展的基本方向产生影响。第三是价值层面。一方面，人工智能大国应该避免通过云服务、预训练模型等向全球南方国家输出隐含某种特定价值观的人工智能系统、塑造新型依附关系；另一方面，应在人工智能未来的发展中，充分体现全球南方国家的文化和价值因素。由于当前的人工智能发展基于人类语言系统，所以严格来说，它在价值层面并不是完全中立的。文明的特征既体现在价值和伦理层面，也体现在思维方式上，这些都内含于一个民族的语言系统之中。未来人工智能发展的一个重要宝库就是非西方国家的语言和文化资源，如果能够把这些资源纳入人工智能的语料库，那么我相信会大大增加人工智能的智慧程度，同时也会为全球南方国家和人类文明的发展提供新的可能性.

Finally, Song says that:

• First, Global South countries must rely on themselves, promoting joint self-strengthening from both the aspects of capacity building and institutional building, breaking the structural asymmetry between the Global South and the Global North in AI global governance, rather than ‘waiting’, ‘depending’, or ‘asking’, pinning hopes on developed countries to proactively treat them equally.

• Second, strengthen the AI capacity building of Global South countries, narrowing the AI divide with developed countries as much as possible. Start from improving AI infrastructure, such as ensuring stable power supply and network connectivity, balancing the deployment of computing-power clusters across different regions of the Global South, expanding the range of use of quality AI products and lowering the cost of AI usage. At the same time, promote South-South educational cooperation in the AI field, jointly build AI training centers, and help Global South countries enhance their AI scientific research capabilities and cultivate cutting-edge AI talent.

• Finally, improve the institutional building of AI global governance and enhance the governance-mechanism discourse power of Global South countries. Fully leverage the UN as well as important multilateral mechanisms such as BRICS and the SCO, coordinate the common interests and governance demands of Global South countries in the AI field, and promote the formation of a Southern agenda in the domain of international AI governance. Global South countries must actively provide important concepts for AI global governance, injecting the consensus concepts of Global South countries into the global AI governance process.

宋国友：首先，全球南方国家要依靠自身，从能力建设和机制建设两方面推动联合自强，打破人工智能全球治理中全球南方和全球北方的结构不对称性，而不能“等”“靠”“要”，寄希望于发达国家主动平等对待。其次，强化全球南方国家人工智能能力建设，尽量缩小与发达国家之间的人工智能鸿沟。从改善人工智能基础设施做起，如确保稳定电力供应和网络链接、在全球南方不同区域平衡布局算力集群、扩大优质人工智能产品使用范围并降低人工智能使用成本等。同时，推动人工智能领域的南南教育合作，共建人工智能培训中心，帮助全球南方国家提升人工智能科研能力、培养人工智能尖端人才。最后，完善人工智能全球治理的机制建设，提高全球南方国家的治理机制话语权。充分借助联合国以及金砖国家、上海合作组织等重要多边机制，协调全球南方国家在人工智能领域的共同利益和治理诉求，推动形成国际人工智能治理领域的南方议程。全球南方国家要积极为人工智能全球治理提供重要理念，把全球南方国家的共识性理念注入全球人工智能治理进程。

There’s an article by the United Front Works Department and the National Ethnic Affairs Commission, discussing the new Law of the People’s Republic of China on Promoting Ethnic Unity and Progress. The law was passed in March and comes into effect from July 1, 2026. The article says that the adoption of the law is a “milestone event in the history of the Party’s united front and ethnic affairs work, and constitutes a major legal measure aimed at forging a strong sense of community for the Chinese nation and advancing the building of the Chinese nation community.”

The first section of the article makes three points regarding the “legal positioning” of the law.

This is an important law that declares the Party’s and the state’s major policies and principles regarding ethnic work. Since the 18th Party Congress, General Secretary Xi Jinping, standing at the strategic height of realising the great rejuvenation of the Chinese nation, has planned ethnic work, clearly put forward the forging of a strong sense of community for the Chinese nation as the main line of the Party’s ethnic work in the new era and the main line of all work in ethnic areas, forming General Secretary Xi Jinping’s important thought on strengthening and improving ethnic work, and promoting historic achievements and historic transformations in China’s cause of ethnic unity and progress. Formulating this law is to transform the major theoretical and practical achievements of the Party’s ethnic work in the new era into the will of the state and elevate them into legal provisions, giving full play to the role of the rule of law in consolidating foundations, stabilizing expectations, and benefiting the long term. This law systematically stipulates the guiding thought, basic principles, main tasks, safeguard measures, and legal responsibilities of the cause of ethnic unity and progress in the new era, solemnly declaring our Party’s confidence and determination to unswervingly follow the correct path of solving ethnic issues with Chinese characteristics.（一）这是一部宣示党和国家关于民族工作大政方针的重要法律。党的十八大以来，习近平总书记站在实现中华民族伟大复兴的战略高度谋划民族工作，鲜明提出把铸牢中华民族共同体意识作为新时代党的民族工作的主线、民族地区各项工作的主线，形成了习近平总书记关于加强和改进民族工作的重要思想，推动我国民族团结进步事业取得历史性成就、发生历史性变革。制定民族团结进步促进法，就是要将新时代党的民族工作取得的重大理论和实践成果转化为国家意志、上升为法律规定，充分发挥法治固根本、稳预期、利长远的作用。该法系统规定了新时代民族团结进步事业的指导思想、基本原则、主要任务、保障措施和法律责任，庄严宣示了我们党坚定不移走中国特色解决民族问题正确道路的信心决心.

This is a promotional law designed to safeguard the fundamental and collective interests of the Chinese nation. T The law closely focuses on ethnic unity as the lifeline, is grounded in the overall perspective of the Chinese nation and the shared interests of all ethnic groups, and prominently emphasizes forging a strong sense of the Chinese nation community and advancing its construction, thereby adjusting ethnic relations across the country. It systematically reviews the formation and development process of the Chinese nation community, focuses on explaining the ‘five commons’ historical view of the Chinese nation, and clearly points out that all ethnic groups have condensed into a community of shared destiny characterized by intermingled bloodlines, shared beliefs, interconnected culture, interdependent economy, and mutual affection (血脉相融、信念相同、文化相通、经济相依、情感相亲), profoundly explaining that the formation and development of the Chinese nation community is what the people desire, the general trend, and a historical inevitability. The legislative intent is clear, explicitly directed toward guiding the people of all ethnic groups to firmly establish the community concept of sharing weal and woe together, sharing honour and disgrace together, sharing life and death together, and sharing destiny together, and to firmly identify with the great motherland, the Chinese nation, Chinese culture, the Communist Party of China, and socialism with Chinese characteristics.（二）这是一部维护中华民族根本利益和整体利益的促进型法律。该法紧扣民族团结这一生命线，立足中华民族整体性视角和各民族一致利益，突出强调铸牢中华民族共同体意识、推进中华民族共同体建设，以此调整全国范围内的民族关系。系统梳理中华民族共同体形成发展历程，重点阐明“五个共同”的中华民族历史观，鲜明指出各民族凝聚成血脉相融、信念相同、文化相通、经济相依、情感相亲的命运共同体，深刻阐明中华民族共同体的形成和发展是人心所向、大势所趋、历史必然。立法意旨鲜明，明确指向引导各族人民牢固树立休戚与共、荣辱与共、生死与共、命运与共的共同体理念，坚定对伟大祖国、中华民族、中华文化、中国共产党、中国特色社会主义的高度认同.

This is a basic law for implementing relevant constitutional provisions, handling ethnic affairs, and carrying out ethnic affairs work. This law is the only newly-formulated law since the 18th Party Congress to have a separately established preamble, fully embodying its overarching and programmatic status as a basic law. This law comprehensively implements the constitutional provisions regarding ‘safeguarding and developing the relations of equality, unity, mutual assistance, and harmony among all ethnic groups’, ‘the state makes every effort to promote the common prosperity of all ethnic groups of the nation,’ ‘realizing the great rejuvenation of the Chinese nation’, and ‘prohibiting discrimination against and oppression of any ethnic group, and prohibiting acts that undermine ethnic unity and create ethnic division.’ It clearly stipulates the rights, obligations, and duty requirements of individual citizens, state organs and public functionaries, people’s congresses at all levels, mass organizations, enterprises and public institutions and social organizations, religious groups, grassroots mass self-governance organizations, the military, and other entities, and is a major landmark achievement in the modernization of the governance system and governance capacity for ethnic affairs in the new era. （三）这是一部实施宪法有关规定、处理民族事务和开展民族工作的基本法律。该法是党的十八大以来唯一一部单设序言的新制定法律，充分体现了统摄性、纲领性的基本法律地位。该法全面贯彻落实宪法关于“维护和发展各民族的平等团结互助和谐关系”“国家尽一切努力，促进全国各民族的共同繁荣”“实现中华民族伟大复兴”“禁止对任何民族的歧视和压迫，禁止破坏民族团结和制造民族分裂的行为”等规定，明确规定了公民个人、国家机关及公职人员、各级人大、群团组织、企事业单位和社会组织、宗教团体、基层群众自治组织、军队等主体的权利义务和职责要求，是新时代民族事务治理体系和治理能力现代化的重大标志性成果.

The second section of the article says that the law “has a complete system, comprehensive content, and rich connotations, reflecting a clear value orientation, prominent characteristics of the times, and a profound institutional logic.” 民族团结进步促进法体系完整、内容全面、内涵丰富，体现了鲜明的价值导向、突出的时代特征、深刻的制度逻辑.

It then makes three points:

Persist in highlighting the main line as the guide. This law runs the main-line requirement of forging a strong sense of the Chinese nation community throughout the entire text, always taking as the primary consideration what is conducive to strengthening the commonality of the Chinese nation and enhancing the sense of the Chinese nation community. The Preamble takes a clear and unequivocal stance, comprehensively delineating the historical trajectory of the formation and development of the Chinese nation, and confirming from a jurisprudential standpoint the main-line positioning of forging a strong sense of community for the Chinese nation. The General Provisions set standards and orientation, and for the first time provide legal interpretations of major foundational concepts such as forging a strong sense of the Chinese nation community and advancing the building of the Chinese nation community, implement the working principle of enhancing commonality while respecting and accommodating differences, and establish norms and benchmarks that anchor to, serve, and safeguard the main line. The measures reflect this fully: in the area of building a shared spiritual homeland, they clearly specify firming up the ‘five identifications’ and guiding the masses of all ethnic groups to firmly establish correct views of the state, of history, of ethnicity, of culture, and of religion; in the area of promoting interaction, exchanges, and integration, they clearly specify improving the social conditions for the masses of all ethnic groups to live together, study together, build together, share together, work together, and enjoy together; in the area of driving common prosperity and development, they clearly specify that policies should be conducive to forging a strong sense of community for the Chinese nation and advancing the construction of the Chinese national community, conducive to safeguarding national unity and opposing separatism, and conducive to improving people’s livelihoods and rallying people’s hearts, thereby indicating the target requirements for ethnic work in the new era and all work in ethnic regions to focus efforts on the main line. （一）坚持突出主线引领。该法将铸牢中华民族共同体意识主线要求贯穿全篇，始终把有利于强化中华民族的共同性、增强中华民族共同体意识作为首要考量。序言旗帜鲜明，全面揭示中华民族形成发展的历史脉络，从法理上确认了铸牢中华民族共同体意识的主线定位。总则立标定向，首次对铸牢中华民族共同体意识、推进中华民族共同体建设等重大基础概念作出法律阐释，落实增进共同性、尊重和包容差异性的工作原则，树立了锚定主线、服务主线、保障主线的准则标尺。举措体现充分，在构筑共有精神家园方面，明确坚定“五个认同”、引导各族群众牢固树立正确的国家观、历史观、民族观、文化观、宗教观；在促进交往交流交融方面，明确完善各族群众共居共学、共建共享、共事共乐的社会条件；在推动共同繁荣发展方面，明确应当有利于铸牢中华民族共同体意识、推进中华民族共同体建设，有利于维护国家统一、反对分裂，有利于改善民生、凝聚人心，指明了新时代民族工作、民族地区各项工作向主线聚焦发力的目标要求.

Persist in combining upholding the correct with innovation. This law adheres to the correct direction, unity of the rule of law, institutional stability, and policy continuity; it stipulates adherence to the comprehensive leadership of the Communist Party of China; it stipulates important principles such as ethnic equality, ethnic unity, and ethnic progress; it stipulates work themes such as joint unity and struggle and common prosperity and development; and it stipulates upholding and improving the system of regional ethnic autonomy. These important provisions consistently adhere to the basic positions, viewpoints, and methods of Marxist ethnic theory, and in one continuous line inherit and carry forward the Party’s basic theories and basic policies on ethnic work. At the same time, this law, with an eye toward the needs of the cause of ethnic unity and progress, emphasises enriching the policy supply, stipulating new requirements such as establishing and highlighting shared and commonly enjoyed symbols of Chinese culture and the image of the Chinese nation, and promoting the construction of China’s independent systems of historical materials, discourse, and theory on the Chinese nation community; it emphasises aligning with new regulatory requirements, fully reflecting relevant content from laws and regulations such as the Patriotic Education Law promulgated in 2023 and the newly revised Law on the National Common Language and Script; it emphasizes a problem-oriented approach, proposing innovative measures for shortcomings in areas such as international communication capacity and coordinated regional development, fully reflecting our Party’s theoretical character of advancing with the times, its people-centered governing philosophy, and its enterprising spirit of pioneering and innovation.（二）坚持守正创新结合。该法坚持正确方向、法治统一、制度稳定、政策延续，规定了坚持中国共产党的全面领导，规定了民族平等、民族团结、民族进步等重要原则，规定了共同团结奋斗、共同繁荣发展等工作主题，规定了坚持和完善民族区域自治制度。这些重要规定，一以贯之坚持了马克思主义民族理论的基本立场、观点、方法，一脉相承赓续了党的民族工作基本理论和基本政策。同时，该法着眼我国民族团结进步事业发展所需，注重丰富政策供给，规定了树立和突出各民族共有共享的中华文化符号和中华民族形象、推动中国自主的中华民族共同体史料体系、话语体系、理论体系建设等新要求；注重衔接新规要求，充分体现2023年颁布的爱国主义教育法、新修订的国家通用语言文字法等法律法规有关内容；注重突出问题导向，就国际传播能力、区域协调发展等存在的短板提出创新举措，充分体现了我们党与时俱进的理论品格、以人为本的执政理念、开拓创新的进取精神.

Persist in giving equal weight to guidance and constraint. As a promotive-type law, this law takes advocacy, guidance, support, and encouragement as important methods; it highlights the promotion and dissemination of the concept of forging a strong sense of the Chinese nation community; it puts forward support measures in various aspects such as cultural integration, social mutual embedding, and economic development; it issues a mobilisation call for the whole Party and the whole society to participate together; it is conducive to giving play to the important functions of value guidance, policy guidance, and behavioral regulation; and it promotes the sense of the Chinese nation community to take root and grow, and to enter deeply into people’s hearts. At the same time, it emphasises the regulatory nature of law, ensuring binding force and enforceability. It sets out legal responsibilities, respectively, for situations where relevant entities fail to fulfill their duties adequately and where organizations and individuals undermine ethnic unity and progress, constructing a multi-entity, multi-level responsibility system. It clearly specifies that for organizing, planning, or carrying out violent terrorist activities, ethnic separatist activities, or religious extremist activities that constitute crimes, criminal responsibility shall be pursued in accordance with law, and it makes provisions on the extraterritorial applicability of the law, further enriching the legal weapons for safeguarding our country’s sovereignty, security, and development interests. This series of provisions combines soft guidance with hard constraint, which is conducive to driving the forging of a strong sense of community for the Chinese nation to be internalized as the common conviction of the masses of all ethnic groups and externalised as the conscious action of all citizens. （三）坚持引导约束并重。该法作为促进型法律，以倡导、引导、支持、鼓励为重要方式，突出铸牢中华民族共同体意识的理念宣介，提出文化融通、社会互嵌、经济发展等各方面的支持措施，发出全党全社会共同参与的动员号令，有利于发挥价值引领、政策引导、行为规范的重要功能，推动中华民族共同体意识落地生根、深入人心。同时，注重法律的规制性，确保具有约束性、执行力。针对相关主体履职不到位、组织和个人破坏民族团结进步的情形分别设置了法律责任，构建了多主体、多层次的责任体系。明确对组织、策划、实施暴力恐怖活动、民族分裂活动或宗教极端活动，构成犯罪的，依法追究刑事责任，并对法律境外适用效力作出规定，进一步丰富了维护我国主权、安全、发展利益的法律武器。这一系列规定，将柔性引导与刚性约束相结合，有利于推动铸牢中华民族共同体意识内化为各族群众的共同信念，外化为全体公民的行动自觉.

The third and final section of the article makes four brief points about implementation.

First, uphold the Party’s leadership. “It is necessary to always take General Secretary Xi Jinping’s important thought on strengthening and improving ethnic work as the fundamental guideline; in promoting the implementation of the law, accurately grasp the relationships between commonality and diversity, between the sense of the Chinese nation community and the consciousness of each ethnic group, between Chinese culture and the cultures of each ethnic group, and between the material and the spiritual, achieving the dialectical unity of commonality and diversity, and the organic combination of ethnic factors and regional factors. 必须始终以习近平总书记关于加强和改进民族工作的重要思想为根本遵循，在推动法律实施中准确把握共同性和差异性的关系、中华民族共同体意识和各民族意识的关系、中华文化和各民族文化的关系、物质和精神的关系，做到共同性和差异性的辩证统一，民族因素和区域因素的有机结合.

Second, strengthen propaganda and interpretation, to enhance awareness about the law, “continuously deepening the whole of society’s cognitive recognition of the law’s core essential meaning and practical value.” It also calls to “promote the law’s publicity and implementation in government organs, work units, communities, schools, and into social organizations, creating a good atmosphere for the law’s implementation.”

Third, advance the governance of ethnic affairs in accordance with law. “Taking the law’s promulgation and implementation as an opportunity, drive all localities, based on their actual conditions, to carry out the revision and improvement of relevant regulations and policies in a timely and steady manner. Whether issuing laws and regulations or policy measures, it is necessary to take whether it is conducive to forging a strong sense of the Chinese nation community as the primary consideration. It is necessary to effectively promote the implementation of the law’s provisions, safeguard measures, and legal responsibilities, and continuously transform institutional advantages and legal advantages into governance efficacy. For various responsible entities that fail to fulfill their duties adequately, and for various organizations and individuals that undermine ethnic unity and progress, handle them in accordance with law. Resolutely oppose all acts that, under the pretext of ethnicity, religion, and human rights, carry out smearing and defamation, containment and suppression, infiltration, and sabotage. （三）依法推进民族事务治理。以法律的颁布实施为契机，推动各地结合实际及时稳妥做好相关法规政策修订完善工作。无论是出台法律法规还是政策措施，都要把是否有利于铸牢中华民族共同体意识作为首要考虑。要切实推动法律规定以及保障措施、法律责任落实，不断将制度优势、法律优势转化为治理效能。对各类责任主体履职不力、各类组织和个人破坏民族团结进步等行为，依法予以处理。坚决反对一切以民族、宗教、人权等借口实施污蔑抹黑、遏制打压、渗透破坏等行为.

Promote the formation of a joint working force. This is essentially about inter-departmental and bureaucratic synergies.