From 26 to 28 July 2025 the World AI Conference and the High-Level Meeting on Global AI Governance were held in Shanghai. More notable than the technology on display were the moves at the level of governance mechanisms.
Two institutional outcomes
First, the proposal to establish a World AI Cooperation Organisation. The Chinese government proposed the organisation, with Shanghai as the initial candidate for its headquarters. If realised, it would be the first international, intergovernmental organisation themed on artificial intelligence, positioned in part to help Global South countries accelerate AI development.
Second, the Global AI Governance Action Plan. Released at the meeting, the document sets out 13 concrete actions, stresses respect for sovereignty and fairness for all, addresses the energy and environmental issues raised by AI, and promotes international cooperation and coordinated governance.
Why mechanisms are harder than rules
Over the past few years the output of global AI governance has been mainly principles and initiatives: the OECD principles, the G7 code of conduct, and relevant UN resolutions. These documents provide a consensus on values, but lack an executing body and a standing mechanism.
The cooperation organisation proposal tries to fill that gap. Its core difficulty is that AI governance involves multiple interests — compute, data, models, talent — and national demands differ widely, so the mechanism must reconcile the right to development with security concerns. See our note on the UN WSIS+20 review.
Developing countries moved to the front
Writing “supporting Global South countries to accelerate AI development” into the organisation’s positioning is a differentiated governance narrative: it stresses capacity building and inclusiveness rather than risk control alone.
That orientation is consistent with China’s statements at other international forums, and it gives Chinese AI companies going global an institutional backdrop. See our notes on the China-ASEAN AI governance workshop and India’s AI governance guidelines.
What it means for companies
First, factor governance-mechanism developments into overseas strategy, since international organisations and multilateral mechanisms may bring new access and standards requirements.
Second, watch the energy and environment theme in the action plan, as green compute may become a new compliance and disclosure dimension.
Third, track the pace of later mechanism-building, since moving from an initiative to a treaty usually takes a long time. See our note on the 2026 WAIC global AI governance guide.
Global AI governance is moving from what to say to who does it. For companies, the meaning is that the external variables of compliance now extend from the text of rules to mechanisms and voice. See our note on the internet information service measures for intelligent services.
This article is based on public information on the 2025 World AI Conference and the High-Level Meeting on Global AI Governance and is for reference only; it does not constitute legal advice.