A new institution, not a new idea
On July 16, 2026, representatives of 29 countries gathered in Shanghai to sign the agreement establishing the World Artificial Intelligence Cooperation Organization (WAICO), described in the founding text as "an independent intergovernmental international organization" headquartered in Shanghai (Xinhua; China's official government portal). Founding members include Russia, Pakistan, Kazakhstan, Indonesia, Laos, Belarus, Serbia, Cuba, Brazil, Venezuela and ten African states. Chinese Foreign Minister Wang Yi signed on Beijing's behalf; UN Secretary-General António Guterres attended the ceremony, held on the eve of the World Artificial Intelligence Conference.
What is missing from that list matters as much as what's on it: the United States, the European Union and its 27 member states, the United Kingdom, Japan and South Korea did not sign.
The case for WAICO, stated fairly
Before dismissing WAICO as a geopolitical vanity project, it's worth taking its stated premise seriously. Existing AI governance venues — the G7 Hiroshima Process, the EU AI Act's extraterritorial reach, the UK/US-anchored AI Safety Summits — were built almost entirely by the handful of countries that host frontier AI labs. Countries without a domestic AI industry have had little say in setting norms that will nonetheless govern which models they're allowed to import, how their data can be used to train systems, and what compliance burdens their startups face. WAICO's charter explicitly frames itself as "open to all sovereign states" with no regime-type screening, and oriented toward "development" and closing the "global capability divide" — a genuine gap. Guterres himself made the same point at the same conference, warning that AI "cannot be governed by a handful of countries or a handful of companies" and that "every nation needs a seat at the table" (UN News). On that diagnosis, Beijing and the UN Secretary-General agree.
Why the diagnosis doesn't justify the cure
Agreeing that the Global South has been under-consulted is not the same as endorsing an institution headquartered in Shanghai, sponsored almost entirely by one government, and signed the same week that government showcased its AI industry at the World AI Conference. WAICO's practical effect will be to formalize a channel through which Chinese open-source models, cloud infrastructure and data-governance norms — permissive of state access in ways the EU's GDPR-adjacent AI Act framework is not — become the default option for the 29 signatories and any state that joins later. That's not neutral standard-setting; it's market-share diplomacy wearing a multilateral label. Nothing in the founding agreement text describes binding technical standards, a dispute mechanism, or funding commitments beyond broad UN Charter language, which suggests the organization's early function is closer to a diplomatic anchor than a regulator (Wikipedia summary of the founding agreement, corroborated by the government portal text).
The deeper problem for pro-innovation policy is fragmentation, not Chinese influence per se. AI developers already navigate the EU AI Act, US state-level rules, China's own algorithm-registration regime, and now a UN track — the Global Dialogue on AI Governance, established by General Assembly resolution and already running its first sessions in Geneva (UN Global Dialogue on AI Governance). A fifth, Beijing-anchored track for 29 countries doesn't add safety; it adds a compliance surface that founders and cross-border deployers now have to model, and it increases the odds that companies serving Global South markets face incompatible rules depending on which bloc a given country aligns with.
The unforced error is the boycott, not the bloc
The more consequential story here isn't that China built an institution — it's that Washington, Brussels, London, Tokyo and Seoul had a UN-anchored, genuinely universal alternative already on the table and chose not to use it as a counter-offer. Analysts have noted the awkward timing: WAICO launched days after a major Western AI lab executive publicly urged the US to "take a lead in global AI regulation" — while the actual US government delegation was absent from the room where 29 countries' worth of Global South AI policy got its first shared institutional home (CIO; Reuters via U.S. News). Sitting out a forum you dislike is a defensible tactic only if you're winning the argument somewhere else. Right now, the US and EU aren't offering the 29 WAICO signatories a comparable package of AI capacity-building, compute access or model licensing terms — so the absence reads less like principled non-participation and more like ceding the floor.
The fix isn't to join WAICO on Beijing's terms. It's to treat the UN Global Dialogue on AI Governance as the legitimate universal venue it already is, show up to it consistently, and pair that engagement with a real capacity-building offer to the Global South — compute credits, open licensing, technical training — so that countries currently defaulting to Shanghai's terms have an actual alternative rather than an empty seat where Washington used to sit.