From 17 to 20 July 2026 the 2026 World AI Conference and High-Level Meeting on Global AI Governance was held in Shanghai. The four-day event was not only a product showcase but a concentrated release of governance frameworks.
For Chinese AI companies, the signals from this conference bear directly on the regulatory environment for going global in the coming years.
A new organisation: the World AI Cooperation Organisation
During the conference, the signing ceremony for the Agreement on the Establishment of the World AI Cooperation Organisation was held in Shanghai, and the organisation’s headquarters were set in Shanghai. It is the world’s first intergovernmental international organisation themed on artificial intelligence.
Its mandate is to support countries of the Global South in accelerating AI development and improving innovation, application and governance capability. For companies, a multilateral organisation in whose creation China participated and took a leading role means that AI cooperation projects aimed at emerging markets may gain clearer policy channels and coordination mechanisms.
Two action plans: ethics governance and cooperative development
The conference released the International AI Ethics Governance Action Plan and the AI Cooperation and Development Action Plan.
The former was guided for release by the Ministry of Industry and Information Technology together with relevant parties, providing a global AI governance action framework that is both principled and operable. The latter sets out actions in eight areas — data, computing power, ecosystem, empowerment, talent, rules, governance and ethics — including quality data supply, inclusive intelligent computing, open-source ecosystem sharing, joint rule and standard building, and security-governance collaboration.
For companies going global, the value of these two documents is that they define the governance language of the “Chinese approach”. In markets such as Southeast Asia, the Middle East and Latin America, understanding and using this language is more likely to win local regulators’ recognition than simply citing European and American frameworks.
Agent interoperability: a new entry point for technology governance
The conference also released the Global Cooperation Initiative on Agent Interoperability. Mutual trust and interoperability between agents are evolving from a technical problem into a governance problem: when one agent calls another on behalf of a user, how responsibility is allocated, how data is authorised and how behaviour is audited all require rule support.
This echoes the approach to agent security in China’s AI Safety Governance Framework 3.0, discussed on our EU AI Act practice page.
Three points for companies going global
First, watch the compliance pathways that multilateral mechanisms bring. Participating in China-led international cooperation projects often requires meeting domestic filing and data-compliance requirements, and may bring a friendlier access environment. See our Multi-Jurisdiction Compliance service page for the existing discussion of multi-jurisdiction adaptation.
Second, bring “governance language” into go-to-market materials. Actively showing adherence to ethics governance and security assessment in materials for overseas regulators and partners is becoming a plus.
Third, watch the responsibility design of agent products. Agent interoperability means cross-entity calls, and product terms need to set responsibility boundaries and data authorisation in advance. See our Commercial Contracts service page.
For support on go-global compliance pathways, cross-border data arrangements and agent product terms, see our Multi-Jurisdiction Compliance and EU AI Act services.
This article is based on the public outcomes of the 2026 World AI Conference and High-Level Meeting on Global AI Governance held in Shanghai from 17 to 20 July 2026. It is for reference only and does not constitute legal advice.