At the 2025 World Artificial Intelligence Conference in Shanghai, China unveiled the Global AI Governance Action Plan (GAGAP). Far from a technical blueprint, GAGAP signals a geopolitical shift that artificial intelligence should be governed less as a laboratory innovation and more as a deployed infrastructure, embedded across industry, governance, and society worldwide. By reframing AI as both digital infrastructure and an international public good to empower economy and society development, GAGAP asserts that large-scale implementation by governments, NGOs, enterprises, and individual citizens, rather than speculative research or frontier technological breakthroughs, will define the next stage of technological agenda.

This doctrine of deployment as progress reframes the global AI debate. The United States’ Winning the Race: America’s AI Action Plan (AAIAP, 2025) stresses innovation primacy and semiconductor dominance; the EU’s AI Continent Action Plan (EU AICAP, 2025) represents a sovereign, values-based ecosystem of data accessibility. China, by contrast, privileges scale, integration, and socio-economic utility. This divergence in models raises fundamental questions, what shapes AI governance in an interoperable ecosystem: frontier innovation, normative regulation, or large-scale deployment. This op-ed argues that China’s approach could accelerate digital inclusion while facilitating equally balanced governance, addressing the digital divide, particularly in the Global South.

Digital infrastructure: A deployment-led pathway to AI innovation

GAGAP is not an isolated initiative but builds on China’s domestic policy trajectory. Since 2015, the Internet Plus initiative has extensively integrated internet technologies into economic and social sectors and afforded China a critical advantage with vast data reserves and extensive real-world digital deployment scenarios (Fan, 2025). This foundation now underpins the AI Plus initiative, which seeks to embed AI across industry, governance, and public services, leveraging the groundwork to drive industrial upgrading and governance innovation (Li, 2024). In this model, deployments generate data and practical insights that fuel further technological refinement and innovation, and sustain further deployment, ensuring China a self-reinforcing cycle. AI is treated not as a theatre of technological competition, but as an engine of innovation derived from practice.

Internationally, GAGAP extends this model outward, advocating AI as shared infrastructure and promoting global exchanges and cooperation. It departs from the US tech focused approach, which underscores technological innovation of lab-based research and development. This prioritisation is encapsulated in AAIAP, which frames around removing barriers to sustain national AI leadership (The White House, 2025). China on the other hand places greater emphasis on adoption across sectors such as smart cities, logistics, healthcare, education, and transportation. This deployment-led approach may appeal to European municipalities or industries facing pressure for affordable and scalable solutions.

Leveraging AI deployment within the Belt and Road Initiative (BRI) reinforces economic empowerment and social development across participating sectors. As BRI used physical connectivity, this approach uses infrastructural integration to popularise immediate application-ready AI achievements, consolidate technological capacity, and address the digital divide (Zhang et. al, 2022). The priority is usability and accessibility, making it especially attractive to Global South nations seeking rapid digital modernisation, enhancing an interoperable AI with compatibility, adaptation, and interconnectivity.

The tension between public good and state power

GAGAP casts AI as a global public good, asserting public sectors should become leaders and pacesetters in AI application and governance, calling for open cooperation, capacity-building, and the reduction of digital divides. It encourages establishing cross-border community and open-source platforms to lower the thresholds of technological innovation and application, and reduce redundant investment in service of humanity, aligning with UN’s 2030 Agenda for Sustainable Development.

Yet these commitments remain at a fledgling stage, and the promises of inclusivity are state-centric and lack operational substance. Domestically, China’s deployment is driven through top-down coordination, guided by sovereignty and controllability (Zeng, 2025). Coastal provinces of China have taken the lead in technological innovation and the growth of AI-driven economies, largely due to favourable policy support. However, this progress produces inconsistencies, as strong central directives coexist with uneven regional capabilities and limited mechanisms for bottom-up feedback (Roberts et. Al, 2021; Wang et al., 2025). Internationally, reciprocal data governance and equitable technology transfer, which is provided by technologically leading countries, are confined to high-level declarations without concrete mechanisms for implementation (Wang et al., 2025). For example, pledges for sustainable AI through green computing and efficiency standards are largely aspirational in the absence of transparent carbon accounting or independent verification. This raises concerns over whether such an infrastructure-led approach genuinely reduces digital inequalities, or reconfigures them under Chinese technological benefits (Züger & Asghari, 2023). It is worth continuing to track how China will reconcile its openness and public interest with the practice of infrastructure diplomacy to achieve interoperability in the digital divide.

Governance interoperability: Connecting a fragmented landscape

GAGAP emphasises the role of industry in formulating technical standards and ethical oversight while building an inclusive multi-stakeholder governance. In practice, establishing unified standards and implementation patterns for an expanding digital infrastructure is difficult. The challenge lies in creating a consistent, inclusive, and fair multilateral AI governance system while aligning with the international community. This is complex because digital development is uneven across fragmented regions, and in differing political contexts (Cath et. Al, 2018).

For the United States, this challenges the assumption that innovation superiority alone ensures global leadership (Hine & Floridi, 2024). By exporting application-ready systems, China offers alternative technological ecosystems that may diverge from US-led standards. How European governance frameworks respond to an infrastructure-led model is not merely external competition, but also internal temptation that municipalities and companies may be drawn to the affordability and efficiency of Chinese AI infrastructure solutions. Such a dynamic embodies a fundamental test for the resilience of Europe’s rights-oriented framework.

The strategic question, therefore, is not merely how to respond to China, but how to engage with a model where technological diffusion is deeply intertwined and shapes a global AI governance. An interoperable AI ecosystem that builds technical and governance bridges between different contexts, will be essential. Enhancing interoperability would allow diverse approaches to coexist: enabling the European risk-based approach to engage with the Chinese implementation scale (Ebers, 2025), the American innovation to connect with global deployment networks, and emerging economies including developmental priorities of the Global South to adopt context-appropriate technologies, without forcing convergence into a single model.

GAGAP shapes a new AI governance logic, regarding deployment as its primary engine and digital infrastructure as a diplomacy instrument (Hamidouche, 2023). By promoting AI as digital infrastructure, China reframes the geopolitics of technology, shifting attention from innovation races to global development. The strategic imperative, therefore, is to secure interoperability underpinned by ethical oversight rather than slide into fragmentation, guiding it away from uncontrolled trajectories towards socially beneficial and globally inclusive development. Only by recognising that AI’s future will be shaped as much by its integration into the fabric of societies as by breakthroughs in research labs can states ensure that deployment advances democratic values, promotes sustainability, and protects collective well-being (Dignam, 2020). Embracing infrastructure-led approach benefits while preventing the reconfiguration of digital divides remains a critical challenge of global AI governance.