World Artificial Intelligence Cooperation Organization Opens in Shanghai, but Its Authority Remains Unproven

The World Artificial Intelligence Cooperation Organization formally launched in Shanghai with 29 founding countries, creating a new center of gravity for global AI governance.

Representatives signed the founding agreement on July 16, 2026. Chinese Foreign Minister Wang Yi signed on behalf of China, while officials from Kazakhstan, Laos, Pakistan, Russia, and Indonesia joined the ceremony.

The organization, commonly shortened to WAICO, will operate as an independent intergovernmental body headquartered in Shanghai. Its stated mission covers international AI cooperation, global governance, safety, fairness, and broader access to technological development.

Yet the signing creates a harder question than its celebratory language suggests. The world already has AI principles, national laws, diplomatic forums, and a United Nations dialogue. What it lacks is agreement about who should set the rules.

WAICO therefore enters a crowded institutional field. Its challenge is not producing another declaration about responsible AI. It must show that a China-hosted organization can turn broad principles into trusted cooperation across political and technical divides.

That tension places WAICO beside, and potentially against, a governance model centered on existing United Nations processes. It also reflects a wider contest between inclusive capacity building and strategic competition over chips, models, standards, and infrastructure.

The World Artificial Intelligence Cooperation Organization Starts With 29 Founding States

The signing turns a Chinese governance proposal into a formal international institution, although many operational details remain undisclosed.

The founding announcement describes WAICO as an independent intergovernmental organization. It says the organization will follow the purposes of the United Nations Charter and apply a human-centered approach.

China says WAICO will promote AI that develops in a beneficial, safe, fair, healthy, and orderly direction. These goals resemble language used by several existing international initiatives, but the organizational form is important.

An intergovernmental organization is established by participating governments through an international agreement. Unlike an industry alliance, it can build permanent governing bodies, membership rules, budgets, programs, and diplomatic processes.

The agreement gives WAICO a headquarters in Shanghai. This makes the city more than an annual host for technology exhibitions and policy conferences. It becomes the administrative base for an institution intended to influence international AI policy.

The 29 signatories became founding member states. The public announcement identifies representatives from Kazakhstan, Laos, Pakistan, Russia, and Indonesia, but it does not publish the complete membership list.

Indonesia’s government separately confirmed that Coordinating Minister for Economic Affairs Airlangga Hartarto traveled to Shanghai to sign the agreement. Its delegation statement described WAICO as a forum for governance, ethical standards, and collaborative AI development.

United Nations Secretary-General António Guterres and representatives of other countries and international organizations attended the ceremony, according to China’s foreign ministry. Attendance, however, does not equal membership or institutional endorsement.

The timing was deliberate. The ceremony took place one day before the 2026 World Artificial Intelligence Conference and High-Level Meeting on Global AI Governance opened in Shanghai.

That conference runs from July 17 through July 20 under the theme “Intelligent Partners, Co-create the Future.” It gives China an immediate diplomatic stage for explaining WAICO’s purpose and attracting additional support.

The launch also follows China’s 2025 proposal for an international AI cooperation organization. That earlier proposal focused heavily on giving developing countries a larger voice in rulemaking and access to AI capabilities.

The July 2026 agreement advances that proposal, but it does not settle the institution’s practical design. The announcement does not explain voting rules, financial commitments, accession requirements, enforcement powers, or relationships with existing organizations.

Those omissions matter because they determine whether WAICO becomes an operating institution or primarily a diplomatic forum. A headquarters and founding agreement create legal structure, but authority depends on what members do next.

Why Shanghai Is Becoming China’s AI Governance Capital

Locating WAICO in Shanghai connects international diplomacy with one of China’s most visible AI industry centers.

Shanghai has hosted the World Artificial Intelligence Conference since 2018. The event brings together government officials, researchers, technology companies, investors, and international organizations.

In recent years, the conference has placed greater emphasis on governance alongside commercial applications. Its program now combines product demonstrations with high-level discussions about safety, standards, development, and access.

China also used the 2025 conference to release a Global Artificial Intelligence Governance Action Plan. That document called for international coordination, shared benefits, capacity building, open-source cooperation, and greater representation for developing economies.

The creation of the World Artificial Intelligence Cooperation Organization gives that policy agenda a permanent institutional address. It also allows China to host meetings, technical programs, and diplomatic negotiations throughout the year.

Shanghai offers practical advantages. The city has universities, research institutions, AI companies, financial infrastructure, and a large international business community. It can connect policy discussions with developers and companies deploying AI systems.

The headquarters also carries geopolitical meaning. Major international institutions often shape the influence of their host countries through staffing, agenda setting, and sustained diplomatic access.

China has argued that existing technology governance systems do not sufficiently represent the Global South. WAICO provides a venue where emerging economies can organize around access to models, computing resources, training, and public-sector applications.

That focus addresses a genuine imbalance. Countries without large cloud platforms, advanced chips, local-language datasets, or specialized researchers face different AI problems from the United States, China, and Europe.

For these governments, governance is not limited to restricting dangerous systems. It also involves obtaining infrastructure, developing skilled workforces, protecting local data, and preventing dependence on foreign vendors.

WAICO can appeal to that wider agenda. It can treat technical capacity as part of governance, instead of separating development assistance from discussions about safety and regulation.

However, hosting the organization in China creates trust questions that its founding members will need to address. Governments outside China’s diplomatic orbit will examine its leadership selection, funding sources, transparency, and access for independent experts.

They will also ask whether WAICO can accommodate competing positions on privacy, human rights, content controls, cybersecurity, intellectual property, and cross-border data transfers.

Shanghai gives the organization resources and visibility. It does not automatically give it legitimacy among governments that approach Chinese technology policy with caution.

That distinction will shape WAICO’s next phase. The institution can become a bridge only if countries with substantially different systems believe they can influence its work.

The Real Contest Is WAICO Versus Fragmented Global Governance

WAICO’s main opponent is not another country or institution, but a fragmented system that produces overlapping principles without consistent implementation.

The international AI policy landscape already includes the United Nations, the OECD, the Group of Seven, the Group of Twenty, regional organizations, national regulators, and technical standards bodies.

These institutions often agree on broad language. They support human oversight, safety, transparency, accountability, fairness, privacy, and cooperation. Their differences become clearer when governments must translate those values into enforceable rules.

The OECD AI Principles, adopted in 2019 and updated in 2024, were the first intergovernmental standard focused on AI. They include five values-based principles and five policy recommendations.

The OECD says 47 jurisdictions adhere to those principles. Governments have used them as a reference for national strategies and risk-management frameworks.

The European Union chose a more prescriptive route. Its AI regulatory framework assigns obligations according to risk and creates specific requirements for providers and users.

The EU AI Act entered into force in August 2024. Most provisions become fully applicable in August 2026, placing implementation almost directly beside WAICO’s launch.

The United States has emphasized innovation, infrastructure, exports, and strategic leadership. The White House’s AI Action Plan contains more than 90 federal policy actions across innovation, infrastructure, and international diplomacy.

That plan explicitly treats AI leadership as an economic and national-security competition. It aims to extend American hardware, models, software, applications, and standards through partnerships with allied countries.

China’s framing emphasizes shared development, sovereign participation, and resistance to technological exclusion. WAICO gives that approach a dedicated institution rather than a temporary diplomatic initiative.

These approaches are not mutually exclusive. A government can adopt OECD principles, regulate high-risk applications, purchase American infrastructure, and participate in a China-hosted cooperation forum.

The problem arrives when obligations conflict. One framework might prioritize model transparency, while another protects commercial secrets. One might require local data controls, while another supports unrestricted cross-border services.

Export restrictions add another layer. Governments cannot promise equal AI access while advanced chips and related manufacturing equipment remain central instruments of strategic policy.

Standards can also become competitive tools. A technical specification that appears neutral can favor certain architectures, vendors, languages, testing methods, or regulatory assumptions.

WAICO’s opportunity lies in coordinating areas where existing frameworks remain disconnected. Those areas include safety testing, multilingual evaluation, incident reporting, research exchanges, public-sector training, and access to computing resources.

Its risk is duplication. If WAICO creates another set of voluntary principles without common projects, reporting mechanisms, or measurable commitments, it will deepen the fragmentation it claims to address.

The decisive test is therefore operational. Can 29 governments use the organization to complete work that they could not complete through existing forums?

Global South Representation Is WAICO’s Strongest Argument

WAICO has the clearest case when it focuses on countries that consume AI technology but rarely control its infrastructure or rules.

Global AI debates often center on the largest model developers. Policy discussions track the behavior of American and Chinese companies, the EU’s regulatory response, and the availability of advanced semiconductors.

That focus can obscure the conditions facing most governments. Many countries rely on imported cloud services, externally trained models, and software designed for languages with far larger digital datasets.

A model can perform well in English while producing unreliable results in local languages. Safety filters can misunderstand regional history, political terminology, cultural practices, or public-health information.

These weaknesses affect real deployments. A government using AI for education, benefits administration, translation, agriculture, or healthcare needs evaluations that reflect local populations.

The same country may lack the computing capacity required to test a large model independently. It may also lack enough specialists to audit vendors, investigate incidents, or negotiate procurement terms.

WAICO can address those gaps through shared testing resources, fellowships, training programs, model evaluations, and technical assistance. It can also support regional datasets under locally defined governance rules.

Such projects would give smaller countries more leverage when purchasing AI services. Shared evaluation methods could help public agencies compare systems without relying exclusively on vendor claims.

The organization could also improve access to safety research. Frontier model assessments require technical expertise and computing resources that remain concentrated in a small number of companies and countries.

Capacity building is already part of the wider United Nations agenda. Guterres has repeatedly argued that every country needs opportunities to shape AI governance, not simply receive systems designed elsewhere.

The Global Digital Compact provides a broad framework for digital cooperation and AI governance. It calls for inclusive participation, scientific understanding, and support for developing countries.

WAICO must explain how its work adds to that framework. The strongest answer would involve delivery: regional laboratories, public benchmarks, training cohorts, shared infrastructure, and transparent funding.

The weakest answer would be another series of meetings where officials repeat principles already found in UN and OECD documents.

Representation also requires more than state participation. AI affects workers, educators, journalists, artists, researchers, children, and marginalized communities. Governments do not always represent those interests completely.

WAICO will need credible channels for universities, civil society, technical communities, and companies. Those participants must have more than ceremonial speaking roles.

Smaller countries will also need meaningful influence inside the organization. If funding or diplomatic weight determines the agenda, the promise of equal participation will quickly lose credibility.

The founding agreement’s public summary emphasizes consultation, joint contribution, and shared benefits. WAICO’s governance rules must show how those ideas work when members disagree.

This is where institutional design becomes substantive. Voting rights, committee leadership, document access, conflict-of-interest policies, and publication standards will reveal who can actually shape decisions.

The Agreement Leaves Its Hardest Questions Unanswered

WAICO has announced its principles and location, but not the mechanisms needed to establish independence, accountability, or technical authority.

The first uncertainty concerns membership. China says 29 countries signed the agreement, yet the initial public announcement does not identify every founding state.

A complete list would help observers assess WAICO’s geographic reach, economic weight, technical capacity, and political diversity. It would also show whether the organization extends beyond countries with close ties to China.

The second uncertainty concerns ratification. Governments often sign international agreements before completing domestic approval procedures. WAICO must clarify when the agreement enters into force and which states have legally ratified it.

The third concerns institutional power. The organization has not publicly detailed whether it can issue standards, conduct evaluations, certify systems, coordinate incident responses, or only provide policy recommendations.

These functions require different resources and governance safeguards. A diplomatic forum can operate with a small secretariat, while a technical assessment body needs laboratories, researchers, secure access, and independent review.

Funding presents another test. WAICO should disclose mandatory contributions, voluntary donations, project financing, and any corporate support. Financial transparency will affect perceptions of institutional independence.

Technical participation creates similar questions. Leading AI companies possess capabilities that governments need, but those companies also have commercial interests in standards and procurement.

An organization dominated by vendors could weaken oversight. An organization that excludes them could lack access to essential expertise and model information.

Human rights provide another difficult area. The founding language refers to a human-centered approach, but governments interpret that principle differently.

Questions about biometric surveillance, political content, automated decisions, military applications, and state access to data will expose those differences. General commitments to safety will not resolve them.

National security could further limit cooperation. Governments may share methods for testing ordinary applications while withholding information about advanced models, cyber capabilities, chips, and military systems.

The founding announcement does not specify whether WAICO covers military AI. Many global governance initiatives limit their scope to civilian or non-military systems because security negotiations require separate channels.

WAICO also needs a clear relationship with the United Nations. Guterres attended the signing ceremony, but the new organization remains institutionally separate from the UN system.

The Global AI Dialogue already gives all governments and stakeholders a UN platform for discussing governance, lessons, and international cooperation.

If WAICO supports that process with research and capacity-building projects, the two can complement each other. If it advances conflicting standards or rival diplomatic blocs, coordination will become harder.

The organization’s headquarters also need safeguards against excessive host-country control. Its credibility will benefit from internationally recruited staff, published procedures, independent audits, and multilingual documentation.

None of these uncertainties proves that WAICO will fail. New organizations rarely disclose every operational detail at a signing ceremony.

However, those details will determine whether other governments treat it as independent. The burden now shifts from diplomatic messaging to institutional evidence.

Three Signals Will Show Whether WAICO Can Influence AI Governance

WAICO’s credibility will depend on transparent membership, concrete technical programs, and coordination with existing international bodies.

The first signal is publication of the founding agreement and full membership list.

Observers need the agreement’s complete text, including its entry-into-force rules, governing bodies, voting process, financing structure, and accession procedures. Governments considering membership need the same information.

A geographically diverse membership expansion would strengthen WAICO’s claim to be a global organization. Growth limited to one diplomatic grouping would instead suggest a narrower coalition.

The second signal is the organization’s first operational program.

A useful early project would address a measurable gap, such as multilingual model evaluation, public-sector procurement standards, AI incident reporting, or researcher training.

The program should identify funding, participants, deadlines, deliverables, and publication requirements. It should also make results accessible outside the founding membership when security considerations allow.

Visible technical work would separate WAICO from forums built mainly around conferences and declarations. It would also create practical incentives for additional governments to join.

The third signal is formal coordination with the United Nations and other standards institutions.

WAICO should clarify whether it will submit research to the UN dialogue, align terminology with international standards, or jointly operate capacity-building programs.

Coordination would strengthen the new organization’s argument that it fills gaps. Competing processes with incompatible reporting requirements would weaken that case.

The next three months should provide early evidence. The 2026 Shanghai conference gives members an opportunity to release documents, announce programs, and explain the organization’s structure.

Personnel appointments will also matter. A secretariat drawn from several regions would support the independence claim. Leadership concentrated in the host country would invite greater scrutiny.

Developers should watch any technical standards or evaluation frameworks that emerge. International benchmarks can shape model documentation, testing practices, and access to public-sector contracts.

Enterprise buyers should monitor procurement guidance, data-transfer principles, and certification proposals. Even voluntary international frameworks can influence contractual requirements across multiple markets.

Knowledge workers and AI users have a direct stake as well. Governance choices affect privacy, content labeling, language support, workplace monitoring, automated decisions, and access to reliable systems.

The World Artificial Intelligence Cooperation Organization has already achieved one significant result. It has persuaded 29 governments to sign an agreement establishing a permanent institution for AI cooperation.

That is more consequential than a conference declaration, but less consequential than an implemented governance system.

WAICO now needs to prove that a Shanghai-based organization can expand participation without becoming another geopolitical bloc. It must also turn broad principles into technical and institutional work that countries can verify.

The central question is no longer whether governments want international AI cooperation. Nearly every major framework says they do.

The question is whether those governments will share authority, resources, information, and responsibility when their strategic interests diverge.

Watch the membership documents, the first funded program, and the relationship with the United Nations. Together, those signals will show whether WAICO becomes a working institution or another layer in an already fragmented system.