Political Tech
China's Shanghai summit signals a strategic pivot from domestic AI regulation to exporting governance norms across the Global South.
NewsOnScale Staff
July 23, 2026
There is a version of the global AI governance race that gets told almost entirely as a transatlantic story — the European Union's risk-based framework on one side, American industry self-regulation and Congressional gridlock on the other. That story is increasingly incomplete.
This week, China hosted a multinational gathering in Shanghai aimed at consolidating an AI governance coalition among countries that have grown skeptical of Western-led digital standards. The summit, organized with explicit diplomatic intent, marks a meaningful escalation in Beijing's effort to do something the EU and US have so far failed to accomplish: turn domestic regulatory posture into international norm-setting.
## What China Is Actually Proposing
China's domestic AI regulatory architecture has matured rapidly over the past two years. Rules covering algorithmic recommendations, deepfakes, and generative AI have all passed through its Cyberspace Administration. Its latest governance push adds provisions on AI ethics, autonomous AI agents, and restrictions on systems designed to pass as human — so-called anthropomorphic AI.
The Shanghai bloc initiative builds on that foundation but projects it outward. By inviting participation from governments in Africa, Southeast Asia, the Middle East, and Latin America, Beijing is offering something concrete: a ready-made governance template, technical cooperation agreements, and the implicit promise of infrastructure investment through channels like Huawei's cloud expansion.
For countries that lack the regulatory bandwidth to build frameworks from scratch, that offer is not trivial.
## The Norm Export Problem
What makes this significant for the AI agent economy specifically is that governance frameworks don't just restrict technology — they shape it. Rules about how AI agents must disclose their non-human status, how they interact with citizens in public services, or how their decisions can be audited directly determine what gets built, by whom, and for whom.
If a coalition of 30 or 40 nations adopts governance norms anchored in Beijing's model, the companies building AI agents for those markets will need to comply with those norms. That has downstream consequences for platform design, data localization, and the competitive landscape for both Western and Chinese AI firms operating globally.
This is the rarely-discussed infrastructure layer of geopolitical AI competition: not just who builds the most powerful models, but who writes the rules that determine how those models behave in public-facing deployments.
## The Western Counter-Position Is Weak
The timing is awkward for democratic AI governance advocates. The US federal AI policy conversation, as detailed in recent Brookings analysis, remains caught between executive action and Congressional inertia. Microsoft's Brad Smith acknowledged this week that the current environment amounts to regulation without clear or complete rules — a vacuum that creates exactly the kind of uncertainty that benefits whichever actor moves with more coherence.
The EU's AI Act is real and enforceable, but its reach is geographic. Brussels can set terms for the European market. It has shown limited capacity to function as an affirmative governance export engine in the way China is now attempting.
## Why This Matters for Civic Transparency
For those tracking AI's role in civic life — government services, public communication, electoral information — the governance bloc question is not abstract. Countries that join China's framework will likely adopt its approach to AI agent disclosure, which tends to favor state-defined standards over independently auditable ones. Accountability mechanisms that civil society organizations and journalists rely on may look very different under those architectures.
The global conversation about AI governance has often been framed as a question of speed: who regulates first, who regulates best. The Shanghai summit suggests the more important question may be about reach: whose rules travel farthest, and who gets to decide what trustworthy AI looks like for the majority of the world's population.
Western governments and multilateral institutions still have room to compete on that dimension. But competing requires showing up with something more than a roadmap.