姚旭：Balancing AI Innovation and Security in China

A type of robot that can be applied in various scenarios, including retail and industry, at the exhibition site of the 2025 World Artificial Intelligence Conference in Shanghai on July 29.(Xinhua)

人工智能正从经济竞争力、国家安全，到社会变革、全球治理等多个维度，重塑现代生活的根基。它并非遥不可及的未来构想，而是定义当下时代的核心力量，这一现实已让全球各国政府、产业界与社会无法忽视。对于中国而言，如何在把握技术快速发展机遇的同时，防范无序发展风险，成为关键议题。复旦发展研究院副研究员、全球人工智能创新治理中心研究员姚旭在《北京周报》（Beijing Review）39期撰文，阐述中国在人工智能治理领域的实践——从国内法规完善、产业生态培育，到国际规则倡议、全球合作推进，展现中国以 “发展与安全并重” 理念平衡人工智能创新与治理的路径，以及为全球人工智能治理贡献的中国方案。

From economic competitiveness to national security, from social transformation to global governance, AI is reshaping the foundations of modern life. AI is not a distant possibility but the defining force of our time, which is a reality that governments, industries and societies worldwide can no longer ignore.

For China, the objective is to seize the opportunities of rapid technological progress while guarding against the risks of unchecked development. This dual commitment, balancing innovation with security—explains why Beijing has moved swiftly and strategically on AI governance. Internationally, China is proposing initiatives and engaging in rule-making. Domestically, it is refining regulations and standards while fostering industrial ecosystems. At the local level, it is experimenting with pilot programs to create exemplary models.

China's AI governance

China's progress in AI is vibrant. Industry, academia and research are increasingly intertwined, producing a surge of influential AI models. By early 2024, China had launched more than 100 large-scale models with over a billion parameters, accounting for more than one-third of global releases, second only to the United States. Behind those numbers lies a deeper story: An expanding network of more than 4,500 AI enterprises, a core industry approaching 600 billion yuan (approximately $84 billion) in value, and a user base of nearly 249 million for registered generative AI services.

The implications are significant. Not long ago, global discussions about AI were dominated by Silicon Valley. Today, however, models such as DeepSeek have demonstrated that China is not only catching up but in some areas setting the pace. For many countries outside the West, particularly in the Global South, this development is more than a technical milestone—it signals the emergence of alternative pathways and partnerships in AI, challenging the notion that advanced innovation is exclusive to a few developed economies.

Yet such growth also brings risks. The risks associated with AI, including bias, disinformation, privacy violations and even systemic threats to stability, are real. Recognizing this, China has taken the unusual step of advancing rules and standards in parallel with innovation.

Domestically, laws such as the Personal Information Protection Law and the Data Security Law create a baseline of accountability. The Interim Measures for the Management of Generative AI Services, which took effect on August 15, 2023, go further, placing red lines around algorithmic manipulation, deep synthesis, and unsafe training practices. These are not abstract regulations; they directly shape how AI companies operate, requiring attention to ethics, transparency and user safety. Complementing these legal safeguards, the National AI Industry Standardization System Construction Guide (2024) sets technical benchmarks in areas such as privacy computing and safety evaluation, combining enforcable laws with more flexible principles and guidelines.

Internationally, China has put forward systemic frameworks for AI governance. In October 2023, it introduced the Global AI Governance Initiative, and on July 1, 2024, the United Nations General Assembly unanimously adopted a resolution titled Enhancing International Cooperation on Capacity-Building of Artificial Intelligence. The resolution, proposed by China and co-sponsored by more than 140 countries, including the U.S., highlights the rights and needs of developing countries. By doing so, China has pushed the issue of narrowing the global "intelligence gap" to the forefront of international debate.

China's governance approach is not confined to national legislation or global forums. On the ground, Shanghai and other cities are actively experimenting with AI development and regulation. Shanghai, which has positioned AI as a pillar of its innovation strategy, has not only incubated leading companies and institutions but also broken new ground in governance. At the 2024 World Artificial Intelligence Conference held last July in Shanghai, the Shanghai Declaration on Global AI Governance was released, emphasizing safety, reliability, controllability and fairness. This provides insights that can be shared at both the national and international levels.

Overall plan and global cooperation

If innovation runs ahead of governance, risks could spiral out of control; if governance is improper, innovation could be stifled. Striking the right balance is a challenge. Here, China's approach of "two hands, both strong"—advancing development while safeguarding security—offers valuable lessons.

China's proposed answer is a full-cycle governance model. This means embedding security considerations from the earliest stages of research and design, applying differentiated regulatory oversight during deployment, and strengthening monitoring and emergency responses during operation. In practice, this would require not only new legal and technical tools but also adaptive regulatory mechanisms capable of keeping pace with fast-moving technologies.

Such an approach is not only about protecting Chinese society. It is about contributing a governance model to the world, one that other countries, particularly those still building their AI capacities, can study, adapt and benefit from.

AI governance is a global challenge that requires the joint participation of all nations. China has consistently emphasized a collective approach, given AI governance is an issue that demands shared solutions. Its engagement with the United Nations, its call to narrow the global intelligence gap and its partnerships across the Global South all reflect a commitment to openness and cooperation. Through these efforts, China positions itself not only as an active participant but also as a leading force in shaping the international governance process.

With developed countries, China is seeking dialogue on AI-related risks and ethics, even in an environment complicated by geopolitical competition. Here, the goal is not to erase differences but to find common ground on issues such as transparency, accountability and responsible innovation.

With the Global South, the agenda is even more pressing. Many developing nations are eager to harness AI for development but lack the resources, expertise, or infrastructure. China's experience, along with its open-source models like DeepSeek and MiniMax, offers pathways for low-cost, accessible solutions. Training programs, joint research, and infrastructure support can help bridge divides and ensure that the benefits of AI are broadly shared rather than concentrated in a few wealthy nations.

By amplifying the voices of developing countries in rule-making, China seeks to promote the formulation of international rules that reflect the concerns of all parties. This approach not only strengthens the participation of Global South countries in the governance of artificial intelligence but also broadens the inclusiveness of the global governance process.

A moment of opportunity

The world cannot afford fragmented approaches that deepen divides or foster mistrust, nor can it tolerate a governance vacuum in which the fastest innovators dictate the rules by default. What is at stake is not only competitiveness but also the fundamental principles of fairness, inclusivity and shared security.

The path ahead will not be easy. Effective governance demands constant adaptation, humility in the face of complexity, and, above all, international cooperation. Yet, if pursued with vision and balance, it can unlock the full promise of AI while keeping its risks in check.

China's actions show that such a balance is possible. By committing to innovation without abandoning responsibility, advancing rules without stifling creativity, and linking national progress to global cooperation, China is presenting a roadmap at a critical moment for the world.

Good governance leads to good intelligence. The present moment—in which AI remains malleable and governance structures are still being shaped—is the best time to act. China's timely leap in AI governance is more than a domestic milestone; it represents a contribution to humanity's shared future. If embraced with openness and cooperation, it holds the potential to deliver greater welfare and brighter prospects for the world.

The author is secretary general of the Center for Global AI Innovative Governance and an associate research fellow at the Development Institute of Fudan University in Shanghai.

Copyedited by G.P. Wilson

Comments toliangxiao@cicgamericas.com

来源 | Beijing Review，原标题为：Balancing innovation and security

作者 | 姚旭