China’s AI Governance Push and the Race to Shape Global AI Rules

On September 12, China’s National Data Administration (NDA) announced that it would develop standards for embodied artificial intelligence (AI) data. The announcement followed a September 10 meeting chaired by NDA head Liu Liehong, where seven key digital companies called for shared public data infrastructure for smart physical machinery and robotics, along with unified data standards. The participants included the startups ModelBest, Infinigence AI, and X Square Robot. While much remains to be seen regarding raw AI development and emerging technologies in general, this announcement is an indicator of a broader trend, with China increasingly building a centralized and coordinated framework for governing how AI is developed and deployed. This is critical because establishing usable rules around AI has significant implications, potentially becoming nearly as vital as the ability to develop the underlying AI models themselves. As a result, safety and governance are becoming strategic components of the global competition over AI. The AI Arms Race AI governance is a critical front in the broader arena of non-kinetic competition between China and the United States. Instead of direct armed conflict, some have described a new “Silent Cold War,” which focuses on topics like AI, semiconductor chips, and foreign investment. AI is especially important because of its dual-use nature, with its capabilities ranging from advising a fifth-grade child on her homework to providing a 99 percent facial verification match on a suspected human trafficker at a border crossing. China’s history of increased AI model exportation, which could result in increased information influence, has been detailed as one of the implications originating from the AI arms race between the U.S. and China. This is particularly important for countries in the Global South, due to the influx of Chinese infrastructure designed to support AI, with many of these efforts being made as part of the Belt and Road and Digital Silk Road Initiatives. If Chinese firms become major providers of AI models and their corresponding cloud services, data centers, and energy grids, then China may also gain influence over technical standards, data practices, content controls, and what defines responsible AI in the first place. This makes AI governance a regulatory issue that must be tackled at an international level rather than at a regional or domestic level. For developing countries deciding which AI systems to adopt, governance frameworks can be a key driver for these nations, with a country selecting an AI provider simultaneously adopting the same standards for data management, content moderation, privacy, and system oversight. AI Governance Consequently, AI governance has become a key frontier topic that must be top-of-mind for policymakers worldwide. AI has and is already being leveraged for nefarious purposes, ranging from human-driven prompts promoting deepfakes and impersonation to AI-generated misinformation. Most recently, U.S.-based Claude maker Anthropic and ChatGPT’s OpenAI faced significant backlash due to key AI safety concerns, with both AI makers disclosing that their own models went rogue in their own capacities. This was compounded by the resignation of a former Anthropic and OpenAI researcher around the same period, with this researcher openly citing these same elements as the reason for his resignation. Therefore, the core regulatory dilemma has become how to regulate AI while still enabling innovation. The U.S. has often approached this challenge through things like voluntary commitments, state-level laws, and a smattering of sector-specific rules. This approach prioritizes flexibility, but at the cost of fragmentation. For example, the U.S. National Institute of Standards and Technology released an AI Risk Management Framework in early 2023, with this currently being revised as part of the White House AI Action Plan released in early 2025. However, this framework remains entirely voluntary. Fragmentation in America’s governance space has remained significant, with some states like Colorado, New York, and California issuing AI-specific legislation, while the vast majority of other states have not. China’s Efforts In stark contrast to America’s fragmented approach, China’s top-down approach has resulted in the creation of a multi-layered regulatory system that covers different parts of the AI lifecycle. This model gives specific regulators concrete levers over AI deployments, even as AI development continues to advance at a blistering pace. The National Data Administration’s approach is the most recent epitome of this approach. It is focusing on developing standards for the data used to train AI for physical machinery and robotics, while also establishing common standards. This is significant because embodied AI depends heavily on data collected in the physical world. Establishing common approaches to collecting, organizing, and evaluating this data could become an important part of building an AI-enabled domestic robotics ecosystem. China has pursued similar governance measures in other efforts in AI. On July 15, the Cyberspace Administration of China (CAC), China’s top internet regulator, brought its Interim Measures for the Administration of Anthropomorphic AI Interaction Services into effect. This law specifically regulates AI systems that simulate human personality traits, with the CAC introducing prohibited content, rules for minors, and highlighting necessary safety assessments for compliance, among others. Additionally, on May 8, the CAC, the National Development and Reform Commission, and the Ministry of Industry and Information Technology jointly released the Implementation Opinions on the Standardized Application and Innovative Development of Intelligent Agents. This marked China’s first comprehensive national policy framework specific to agentic AI, introducing tiered risk governance, priority scenarios, and “off-switches” for exigent circumstances. The rapid and cohesive pace of this legislation highlights China’s desire to govern AI in a more stringent and systematic manner. Chinese regulators are already connecting policy, technical standards, data infrastructure, and implementation. These efforts could increase the amount of Chinese-built AI systems used on a worldwide scale, particularly as Beijing increasingly positions its systems as usable and trusted due to its leadership in AI governance. For many, the AI race seems to be determined by who will create the fastest and most advanced AI system. However, China is demonstrating that this race is instead being decided by which country creates an environment in which AI can be trusted, deployed, and adopted at scale. China’s growing advantage is increasingly visible in its ability to connect regulations, policies, technical standards, and enforcement. As Chinese AI systems and infrastructure become more prevalent internationally, establishing clear governance practices could enable China increased influence over how AI is regulated and deployed on an international level. If America continues to rely on its fragmented and voluntary approach, it risks enabling China to play a larger role in setting the rules that shape the future of the world’s AI economy.