Russia-China AI Alliance Launches 29-Nation Standards Body in Shanghai

For AI researchers and engineers, WAICO’s 'people-centered' approach may translate into mandatory transparency reports, algorithm audits, and potentially state-required backdoors in the name of security. The recent Chinese accusation against Claude Code is a warning sign: technical standards here won't be just academic papers, they could become hard requirements.

On July 16, 2026, Russia, China, and 27 other nations signed an agreement to establish the World Artificial Intelligence Cooperation Organization (WAICO), a new intergovernmental body headquartered in Shanghai. The ceremony, attended by UN Secretary-General Antonio Guterres, marks a decisive moment in the global contest to define the principles governing artificial intelligence—a technology that will shape economic, military, and social power for decades. With 29 founding members spanning Africa, Asia, and Latin America, WAICO immediately positions itself as a counterweight to Western-led governance efforts like the OECD AI Principles, the EU AI Act, and the US executive orders on AI.

“

The recent Chinese accusation that Anthropic’s Claude Code contains “security backdoor vulnerabilities” illustrates how technical standards can be weaponized.

The organization’s charter, as reported by Xinhua, promotes a “people-centered approach” to AI, pledging to make the technology safe, fair, and beneficial to all humanity. This language, superficially universal, echoes the human-centric rhetoric used by the West but may carry fundamentally different operational meaning. In the Russian and Chinese context, “people-centered” governance often emphasizes state authority and social stability over individual rights, a nuance that will shape WAICO’s eventual standards and certification processes. Russian Deputy Prime Minister Dmitry Grigorenko, who co-signed the agreement, explicitly called for “transparent rules governing extraterritorial technologies,” signaling that WAICO may seek to impose its standards on AI systems developed outside its member states.

The geopolitical backdrop is one of fragmentation. The EU AI Act, finalized in March 2024, created a risk-based regulatory pyramid; the US has pursued voluntary agreements and sector-specific guidance; and China has enforced strict AI controls under its internet sovereignty doctrine. WAICO arrives at a time when multilateral AI governance is already strained—the 2023 Bletchley Park and 2024 Seoul safety summits produced voluntary commitments but no binding treaty. By securing UN Secretary-General attendance, WAICO gains immediate multilateral legitimacy, potentially influencing ISO standards and other international technical bodies. Its large membership, including 10 African and 12 Asian nations, gives it a voting bloc that could steer global norm-setting away from the transatlantic axis.

For multinational AI companies—from Anthropic to OpenAI to Google DeepMind—WAICO introduces a new compliance frontier. The recent Chinese accusation that Anthropic’s Claude Code contains “security backdoor vulnerabilities” illustrates how technical standards can be weaponized. WAICO could develop a certification regime that requires algorithm audits, transparency reports, and even code submission under the guise of safety, effectively forcing Western companies to choose between market access and protecting proprietary technology. This recalls data-localization battles that reshaped cloud computing; AI could face similar balkanization.