From 6 to 9 November 2025 the World Internet Conference Wuzhen Summit was held. Two research outputs directly relevant to AI governance were released: the Global AI Standards Development Report and Building a Global AI Safety and Governance System for the Common Good of Humanity.
The core finding: fragmentation is worsening
The second report makes a notable judgment: fragmentation in global AI governance is intensifying. Differences across jurisdictions in risk tiering, prohibitions, transparency duties and data-flow rules make it hard for companies to cover multiple markets with one compliance plan.
The report recommends focusing on four key themes: standards, safety, inclusion and trust. “Standards” is placed first, indicating that the report treats interoperability of technical standards as the foundation of governance coordination — without common standards, certification, testing and assessment cannot be mutually recognised.
Why fragmentation matters to companies
Fragmentation is not abstract; it translates directly into cost. The same generative AI product may be classified as high-risk in the EU and require conformity assessment, need only a filing in Southeast Asia, and face no clear rules in the Middle East. Companies must prepare different documentation, assessments and liability arrangements for each market.
This is also why the report lists “inclusion” as a key theme: if governance frameworks reflect only the concerns of a few advanced economies, emerging markets and SMEs are excluded from rule-making, ultimately undermining the universality of the rules. See our multi-jurisdiction service.
Echoing China’s existing positions
The summit report’s orientation is consistent with the “AI for good” and “people-centred” framing China has long advocated internationally. The China–ASEAN AI governance workshop held in early 2026 continued a similar approach, emphasising regional cooperation and capacity building. See our notes on the China–ASEAN AI governance workshop and the WAIC global AI governance review.
How companies should respond
First, build standard interoperability into product planning. If a product targets multiple markets, consider at the design stage whether it can meet the common core of several standards, leaving configuration room for differences.
Second, maintain a regional rule map. Record prohibitions, high-risk lists, transparency duties and data rules by jurisdiction, avoiding a one-size-fits-all approach. See our multi-jurisdiction service.
Third, watch the certainty created by regional cooperation. Regional governance dialogue often produces a common baseline first, and early adaptation can yield an advantage. See our note on the WIC Asia-Pacific summit.
The value of the Wuzhen report is not new rules but putting the real difficulty of fragmentation on the table. For companies, acknowledging fragmentation and preparing tools for it is more realistic than waiting for a single global rulebook.
This article is based on the Global AI Standards Development Report and Building a Global AI Safety and Governance System for the Common Good of Humanity released at the 2025 Wuzhen Summit, and public reporting. It is for general reference only and does not constitute legal advice.