On July 16, in Shanghai, 29 states signed (external link) the constitutive agreement of a new international organization: WAICO, the acronym for the World Artificial Intelligence Cooperation Organization.
One of WAICO's primary objectives is to foster international scientific and technological cooperation in artificial intelligence (AI), particularly in the development (external link), access, and application of AI technologies and services in a secure, ethical manner and for civilian purposes.
As highlighted in a joint press release (external link) by the Brazilian Ministries of Management and Innovation in Public Services, Foreign Affairs, and Science, Technology and Innovation, WAICO will also promote cost-free open-source ecosystems, as stipulated (external link) in Article 3 of its constitutive agreement, and will coordinate with relevant international processes and organizations, particularly within the framework of the United Nations.
The headquarters of the latest intergovernmental initiative for AI governance will be located in Shanghai, and its founding members include developing countries such as China, Brazil, Russia, Kazakhstan, Indonesia, South Africa, Malaysia, Kenya, Pakistan, and Venezuela, among others.
Yet it is also essential to note those that chose not to join WAICO, such as the United States (US), Japan, India, the United Kingdom, and European Union members including France, Germany, and Spain. This does not mean that global AI governance is not a priority for these countries, but rather that there is an ongoing contest between competing models.
The Geopolitics of AI
Under intergovernmental negotiation for more than a year, WAICO was formally announced by Chinese President Xi Jinping just one day before the 2026 edition of the World Artificial Intelligence Conference (WAIC), also held in Shanghai.
There, the Chinese company Moonshot AI announced (external link) Kimi K3, a large language model (LLM) with 2.8 trillion open-weight parameters—a significant contrast to the models of US companies such as Google, OpenAI, and Anthropic, which feature closed weights and therefore cannot be downloaded or modified by users.
During the WAIC, Xi demonstrated (external link) the centrality of AI governance for China, stating that "it is necessary to work together on the development of AI, so that its advancement does not become a solitary endeavor by a single country, but rather a symphony of international cooperation." He further announced that China will create, over the next five years, five thousand AI training and specialization opportunities for professionals from WAICO member states and from international organizations supporting the initiative, such as the African Union, the Arab League, and the Shanghai Cooperation Organisation, among others.
In his address, the Chinese president also made an appeal (external link) to address the potential risks associated with AI and ensure that it "always remains under human control," thereby steering clear of models that train autonomously to create systems beyond human oversight (artificial general intelligence). Xi also stated that countries must oppose "stretching the concept of national security in the field of AI, or placing the security of one country above that of others."
Although he did not single out any country by name, this was a critique of the various measures adopted in recent years by the US and its allies to restrict China's access to cutting-edge technologies and AI. By spearheading the creation of WAICO, the Chinese state makes clear that digital technologies and AI have become a crucial pillar of its industrial policy, as already outlined in its 2023 Global AI Governance Initiative (external link), and that its goal is to exert increasing influence over the establishment of the principles, rules, and standards that will govern global AI in the coming years.
It is true that regional and international initiatives on AI governance are already numerous. However, they have advanced primarily under the leadership of developed countries, as seen in the G7's Hiroshima AI Process (external link) and the AI Principles (external link) of the Organisation for Economic Co-operation and Development (OECD).
In this landscape, the establishment of an international organization dedicated to AI cooperation and comprised of developing countries—or the Global South—could radically (external link) alter international dynamics, leading to more robust and permanent institutional commitments.
Washington's Reaction
In light of the strong accession to WAICO, the US State Department is preparing what amounts to a veritable ultimatum (external link) for several countries: they must pick a side in the "AI race" with China, and should they choose that path, they will be excluded from the US-led coalition, Pax Silica.
Established in December 2025, Pax Silica is a US initiative (external link) designed to build a "secure, resilient, and innovation-driven technology ecosystem." To that end, the US and 23 other states, alongside the European Union, agreed to "reduce coercive dependencies, build partnerships to secure global technology supply chains, address opportunities and vulnerabilities in the AI supply chain, and explore joint investments, in addition to protecting sensitive technologies and building trusted digital infrastructure."
Pax Silica is therefore heavily focused on access to critical supplies and minerals essential for the production of chips and semiconductors—that is, part of the foundational infrastructure of digital technologies and AI. WAICO, in turn, is geared more toward the formulation of principles, norms, and standards that will influence the domestic legislation of its members. Yet one dimension cannot be conceived without the other. Consequently, from the US perspective, countries that have joined Pax Silica cannot join China-backed WAICO without undermining its foundations and their very alignment with the US.
According to this US rationale, Kazakhstan (external link), a potential source of critical minerals that joined both Pax Silica and WAICO, must make a strategic choice, as it will not be able to simultaneously deploy US-origin infrastructure, such as Nvidia chips, and Chinese-origin infrastructure, such as Huawei chips. Companies in Saudi Arabia (external link) and the United Arab Emirates (external link) are already adhering to this directive, committing, for instance, not to use Huawei hardware.
The geopolitics of AI and the contest over its governance are thus laid bare.
Moving forward, it will be essential to monitor WAICO's initial steps to evaluate its contribution toward narrowing inequalities in the production of and access to increasingly essential technologies—issues that Brazil, as a founding member of the organization, actively champions through its foreign policy.
This article, “Como funcionará e quais os desafios da WAICO, a Organização Mundial para Cooperação em Inteligência Artificial,” by Rúbia Marcussi Pontes, was originally published in The Conversation and is licensed under Creative Commons (CC BY-ND 4.0).