Illustration: Chen Xia/GT

The Australian think tank Lowy Institute recently published an article entitled "The quiet regionalization of AI governance," noting that regional organizations such as ASEAN are promoting greater coordination among member states on artificial intelligence (AI) governance rules in a bid to secure greater autonomy in the AI field. The article describes this approach as a "third path" in global AI governance.

This observation highlights an emerging trend in global AI governance. An increasing number of countries are unwilling to remain on the sidelines of AI governance. Instead, they are seeking to participate in shaping AI governance rules through various means, including regional cooperation, while expanding opportunities and space for their own AI development.

Historical experience suggests that a "third path" often emerges when competition between major powers in a particular field is at its most intense. However, such an approach tends to grow and gain momentum when major-power rivalry reaches a certain plateau. Today, the conditions for this development appear to be taking shape. 

Since the beginning of 2026, China and the US have found increasing common ground on preventing AI safety risks, including an agreement to establish a dialogue on AI. This area is becoming an important component of the efforts by the two countries to build a constructive relationship of strategic stability based on respect, fairness and reciprocity.

Clearly, maintaining China-US dialogue on AI can reduce miscalculations and create a stable environment for technological cooperation among third-party countries. Regional organizations can systematically raise specific issues of common concern within their regions, reflect the shared aspirations of the majority of countries, and advance the agenda for cooperation on global AI governance. In this sense, major-power dialogue and regional cooperation can each play their respective roles while complementing one another.

As China and the US strengthen communication and dialogue on global AI governance, cooperation among countries at the regional level may also reach a new stage. 

First, countries can explore ways to integrate different governance approaches and reduce future compliance costs. Second, they can strengthen safety capabilities and emergency response mechanisms. Through regional cooperation, countries can improve their ability to identify security vulnerabilities, promote AI safety education, and establish more authoritative and effective mechanisms for reporting safety incidents. Third, they can promote multilateral cooperation on local challenges. For example, in addressing problems such as AI-enabled fraud and algorithmic pollution, small and medium-sized countries need access to more advanced technological support to improve their response strategies. Chinese and US tech companies can engage in healthy competition by drawing on their respective technological and service advantages, while also pursuing cooperation based on their individual strengths.

Of course, in exploring a "third path," regional governance models must also guard against becoming closed systems, creating new hierarchies and divisions, or contributing to the fragmentation of global AI governance. If rules developed through regional cooperation become excessively stringent, they could create unnecessary market barriers, making it difficult for countries within those regions to participate in global AI development. 

Meanwhile, if individual countries within a region unilaterally dominate the formulation of AI governance rules and set excessively high entry thresholds, squeezing the development space of other countries, regional cooperation could instead produce negative consequences. Furthermore, a lack of communication and cooperation between different regions, coupled with competition among them, could further fragment the global AI governance system.

Therefore, global multilateral platforms such as the UN have a crucial role to play in promoting regulatory coordination and exchanges of experience among countries from different regions. Such platforms can help transform regional experiments into broader governance consensus.

Countries participating in the regionalization of AI governance should maintain transparency and openness, provide sufficient transition periods, and remain open to foreign companies that comply with relevant rules. While respecting other countries' independent choices, they should also help third-party countries strengthen their capacity to address challenges and exercise regulatory oversight, while fostering domestic AI safety industry ecosystems.

Ultimately, the "third path" should not be understood as an attempt to build an entirely separate system, but rather as a necessary complement to the existing "main routes" of global AI governance. A network with more interconnected routes is less vulnerable to congestion at any single point. Similarly, regional cooperation can serve as a bridge connecting different AI governance systems, enabling more countries to participate in this important agenda that humanity must collectively address.

The author is deputy director and a research fellow at the Institute of American Studies, China Institutes of Contemporary International Relations. opinion@globaltimes.com.cn