By Konstantinos D. Magliveras, Professor of Public International Law at the University of the Aegean, Greece
I. Introduction
Presently, there is so much discussion on artificial intelligence (AI), which is logical as it is the one area where technology moves by leaps and bounds, not to mention the many hundreds of billions of USD in investments and the sky-rocketing stock capitalization of companies, which, directly or indirectly, are engaged in the race towards AI domination. Increasingly, this race takes the form of straightforward antagonism between the USA and China, reminiscent of the antagonism between the former and the USSR during the Cold War. In instances of scientific and technical inter-state competition what matters is not always which country has the most funds to spend. What also matters is which country (or group of countries) can build it faster and cheaper, have more leverage and be more influential on the international plane, exercise more soft power in international relations, etc.
To that extent, becoming the world’s norms-setter in the deployment, application and regulation of AI is no small feat. On 16 July 2026, at the Chinese government’s invitation, 29 states[1] with the presence of the UN Secretary General signed in Shanghai the Agreement establishing the World Artificial Intelligence Cooperation Organization (WAICO); incidentally, it will be headquartered in Shanghai. The following day, President Jinping opened the World Artificial Intelligence Conference (WAIC), attended by representatives from over 100 countries and held in conjunction with a High-Level Meeting on Global Governance of Artificial Intelligence.[2]
II. The purposes and functions of WAICO and the relationship to the UN
According to Article 2 of the Agreement, WAICO has the following purposes: (a) develop and apply AI with a human-centric approach to be available to all; (b) expand international cooperation on AI, and jointly establish an open, fair, just and non-discriminatory environment for its development; (c) bridge the AI divide by promoting greater accessibility of AI technologies and services; and (d) collectively address AI risks, challenges and impacts while promoting a safe, secure, ethical and trustworthy AI. While Article 2 does not provide that WAICO shall only engage in the civilian domain of AI, this is expressly stated in Article 3, the clause elaborating on its purposes. Article 3 also sets out the following WAICO functions, which, arguably, overlap with its purposes: (a) build a platform to match supply-demand among Member States; (b) enhance international cooperation on Al capacity building; (c) align and coordinate AI development strategies, governance rules and technical standards, build policy consensus, promote interoperability, and share best practices; (d) strengthen cooperation on AI science and technology, and encourage joint efforts to develop an open-source ecosystem; (e) promote global AI governance as well as convergence with the process within the U.N. framework; and (f) cooperate with other international organizations.
The reference to the ‘process within the U.N. framework’ is noteworthy. However, in terms of AI regulation, little of real substance has been happening at UN level. Excluding General Assembly (UNGA) resolutions on AI use in military field,[3] the other initiatives include UNGA resolutions on AI and sustainable development, now adopted at each yearly Session;[4] the Pact for the Future approved as Resolution 79/1 on 22 September 2024, including the annex on Global Digital Compact; and Resolution 79/325 setting out the terms of reference of the two entities that the UNGA set up to deliberate AI issues.[5] One of them, the Global Dialogue on AI Governance held its first session in July 2026. These UN initiatives cannot lead to hard law AI regulation, and there is no discussion on producing a treaty regulating AI.
This state of affairs raises significantly the importance of institutional projects like WAICO, which, while multilateral in nature, is a project devised and so far executed by one state, China. It has been suggested that China will use the work of WAICO to influence UN discussions.[6] Even though this remains to be seen, were WAICO successful in producing hard law norms, rules and procedures for AI regulation (they could take the form of legally binding principles and guidelines – no need for a treaty), could the UN not involve WAICO, as an international organisation as opposed to its individual Member States, in the deliberations before the Global Dialogue on AI Governance? Moreover, were the Independent International Scientific Panel on AI tasked with drafting a legal instrument on AI regulation, how could the work of WAICO not be one of the cornerstones?
III. The institutional structure of WAICO
Article 5(1) of the Agreement provides for only two organs, a Council and a Secretariat. Pursuant to Article 6(1), the Council shall be the decision-making body, headed by a Chairperson, who shall be elected by the Council’s annual plenary session ‘with due regard to equitable geographical distribution’. No further details are given about the authority, responsibilities and functions of the Chairperson. Presumably, these and all other required arrangements will be laid down in the Council’s rules of procedure.
Article 7 of the Agreement may pose some problems in the early stages of WAICO’s operation. It stipulates that decisions will be made at Council’s plenary sessions, where quorum is set at a minimum of two-thirds of all Member States, and the mode of decision is consensus. If consensus ‘cannot be reached after every effort has been exhausted’, draft decisions will require a two-thirds majority of the members present and voting. Assuming that WAICO operates with all the 29 original signatories, the positive vote of any 13 members will carry the decision. Here lies a potential problem: under Article 15, the Agreement shall enter into force with only three ratifications. Therefore, WAICO will start functioning with only three Member States; it cannot be known when the other 26 original signatories will ratify it. In the three Member States scenario, the quorum will be two members and decisions can be taken with only one member in favour! Were this to happen, WAICO would be transformed into a bilateral arrangement defying the whole purpose of setting up an international organisation.
Presumably, the organisers’ prospect is for all original signatories to ratify the Agreement (almost) simultaneously and as soon as possible, while a large number of third states will agree to accede to the Agreement (Article 10(3)) within a relatively short period of time. There is no guarantee that either prospect will materialise. Therefore, setting the required number of ratifications to the minimum possible number of three should guarantee an early commencement of operations but, if the number of ratifications / accessions fails to increase substantially, the future of WAICO will be in the balance.
IV. Financing WAICO
In such an eventuality, some members may seriously consider withdrawing from WAICO, which is provided for in Article 17. In particular, withdrawal from the Agreement and from WAICO requires a written notification and the setting of a date, which may not be shorter than of one year. As is customary with the constitutive instruments of international organisations, Article 17 clarifies that withdrawal does not affect obligations arising out of the Agreement. One such obligation is to have settled in full the annual contributions to the budget. This brings up the question of financing WAICO, which is regulated in Article 9 of the Agreement, a clause with several interesting aspects. First, the variety of the sources of financing: not only the Member States’ annual contributions but also contributions from governments (presumably not the governments of Member States), international organisations, corporations, individuals and from other (not specified) entities. Accepting contributions or assistance from these sources is subject to two conditions: (a) the Council must approve it and (b) they may not in any way prejudice, restrict, distort, or alter WAICO purposes, objectives or functions. Second, financing may also take the form of ‘lawful revenue generated from cooperative projects’, a provision that presumably means that WAICO may provide paid services to Member States, to non-members, to groups of states or even to other international organisations. Third, the Council will set Member States’ annual contributions according to an agreed scale of assessment, based on their contribution classes in the system of the Universal Postal Union (UPU). Member States’ economic development level and their capacity to pay could also be taken into account. Again, details are not given in the Agreement, but they should be elaborated when the financial regulations are adopted by the Council (Article 9(1)).
It is not immediately clear why the model of UPU financing was chosen. In a nutshell, it comprises three ‘pillars’: Pillar 1 covering contributions from member countries to the regular UPU budget; Pillar 2 covering contributions to the affiliated Cooperatives[7] in exchange for their services; and Pillar 3 covering fees for chargeable products and services and voluntary contributions for extra-budgetary programmes.[8] As regards Pillar 1, it comprises a system of ten contribution classes with an additional class for the least developed members.
V. Conclusions
According to Article 10 of the Agreement, membership of WAICO is open to all states. No criteria and no requirements are mentioned. The Council will approve membership applications. Evidently, several states or groups of states will never be interested in joining, if for no other reason because they already have a mandatory regulatory regime in place, e.g. the European Union, which in 2024 adopted its own ‘AI Act’, Regulation (EU) 2024/1689. Moreover, the Council of Europe concluded in September 2024 the Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law, which, however, has been signed by less than half of the eligible states and ratified by none.[9] The OECD adopted the Recommendation on Artificial Intelligence in 2019, revised in 2024. But the promoters of WAICO are arguably not concerned with Western states and those participating in the Western international organisations. They are geared towards the states of Asia and Africa, which for the last several decades have collectively dominated the rhetoric at the UNGA and other global meetings, and possibly the countries of South and Latin America. It has also been suggested that WAICO could be seen as offshoot of the BRICS group of countries and their 2025 Statement on the Global Governance of Artificial Intelligence, with China taking the initiative to promote it as it is leading the race towards AI supremacy.[10] Notwithstanding these considerations, it is clear that the development of AI, which mostly takes place in the context of private companies and not at the level of states, has to be regulated not at domestic level or regional level or even continental level but at global level. This is a role that should have been fulfilled by the UN (Articles 1 and 13 of the Charter). With the UN in tatters and lost in myriad meaningless (but highly paid) discussions, no wonder why newcomer stakeholders will try to usurp it.-
Photo by Igor Omilaev on Unsplash
[1] Including the Russian Federation, Brazil, the Republic of South Africa, Kenya, Serbia and most ASEAN members.
[2] See Chair’s Statement for both events: https://www.mfa.gov.cn/mfa_eng/xw/zyxw/202607/t20260717_11984715.html. WAIC has taken place in China every July since 2018.
[3] UNGA Resolution 79/239 (2014), Resolution 80/58 (2025) and Resolution 80/23 (2025).
[4] See, inter alia, UNGA Resolution 78/132, 19 December 2023; Resolution 78/265, 21 March 2024; Resolution 78/265, 21 March 2024; Resolution 79/334, 5 September 2025; Resolution 79/322, 25 July 2025. See also Resolution 80/215, ‘Promotion and protection of human rights in the context of digital technologies’, 18 December 2025.
[5] UNGA Resolution 79/325, ‘Terms of reference and modalities for the establishment and functioning of the Independent International Scientific Panel on Artificial Intelligence and the Global Dialogue on Artificial Intelligence Governance, 26 August 2025.
[6] See A. Basu, ‘China’s Pivot on Global AI’, Carnegie Endowment for International Peace, 21 May 2026 <https://carnegieendowment.org/research/2026/05/chinas-pivot-on-global-ai>.
[7] These are subsidiary organs, which, in addition to the UPU’s governing bodies, serve specific roles within the UPU and are financed by Member States; see <https://www.upu.int/en/universal-postal-union/about-upu/cooperatives-boards>. Under Articles 5(3) and 6(1), the WAICO Council may establish subsidiary bodies as necessary.
[8] See <https://www.upu.int/en/universal-postal-union/about-upu/finance>.
[9] See <https://www.coe.int/en/web/Conventions/full-list/?module=signatures-by-treaty&treatynum=225>.
[10] See A. Sytnik, ‘WAICO & BRICS: What Does the New Global AI Organization Mean for the Group?’, BRICS Expert Council – Russia, 18 July 2026 <https://bricscouncil.ru/en/analytics/waico-and-brics-chto-novaya-vsemirnaya-organizatsiya-v-sfere-ii-oznachaet-dlya-obyedineniya>.
International Law Blog 