APEC digital and AI ministers issued the Chengdu Statement on July 23, endorsing open-source development while emphasizing security, data protection and intellectual property rights. CNBC reports that the 21-member forum, which includes the United States and China, also backed models and projects using strong security assurance during development and deployment. APEC's official release lists investment and open source for the digital ecosystem among five consensus areas.
APEC digital and AI ministers issued a joint statement in Chengdu, China, on July 23 that endorses open-source development while emphasizing security, data protection and intellectual property rights. CNBC reports that the statement, agreed by the forum's 21 member economies including the United States and China, supports models and projects that employ strong security assurance through development and deployment.
The meeting was chaired by Li Lecheng, China's minister of industry and information technology. According to the original CNBC report, Li described the statement as the first APEC ministerial-level statement to include cooperation on open source.
Five areas of consensus
APEC's official release identifies five areas covered by the ministerial consensus:
- •Universal and meaningful digital connectivity
- •Digital technologies and AI for economic growth
- •Digital literacy and skills enhancement for all
- •A trusted digital environment
- •Investment and open source for the digital ecosystem
The full Chengdu Statement places the meeting within the APEC Artificial Intelligence Initiative for 2026-2030 and acknowledges that member economies take different approaches to digital and AI policy. Its introductory language also calls for digital and AI deployment that supports long-term regional economic growth.
Open development paired with security language
The notable feature of the statement is the explicit pairing of open-source AI with safeguards. CNBC reported that the text calls for respect for security, data protection and intellectual property rights, rather than presenting open development as separate from governance requirements.
CNBC framed the agreement against a backdrop of tighter controls on domestic AI capabilities in both Washington and Beijing. The outlet also contrasted downloadable models from Chinese companies, including DeepSeek and GLM 5.2, with closed, paid model access offered by companies such as Anthropic. Those company-level comparisons are context from CNBC, not requirements imposed by the APEC statement.
Winston Ma, an adjunct professor of law at New York University, told CNBC that the statement's language on open-source ecosystems and compute infrastructure reflected a regional shift toward open-weight systems combined with state-coordinated energy, telecommunications and digital infrastructure. That is an outside interpretation of the agreement, rather than a commitment attributed to every APEC economy.
What remains operationally unresolved
APEC's press release identifies consensus areas but does not describe implementation details.
For ML teams, agreements that join open development with security assurance often make provenance, vulnerability handling, model evaluation and deployment controls more consequential procurement and governance questions. The Chengdu language pairs open-source cooperation with security assurance, data protection and intellectual property rights.
Key Points #
- 1APEC's 21 economies endorsed open-source AI cooperation while explicitly pairing it with security, data protection and intellectual property protections.
- 2The Chengdu Statement makes investment and open source one of five ministerial consensus areas for regional digital and AI policy.
- 3The statement pairs open-source cooperation with security assurance, data protection and intellectual property rights.
Scoring Rationale #
The statement brings the United States, China and other APEC economies into a shared ministerial position on open-source AI and security assurance. It is relevant to practitioners tracking cross-border AI governance and open-model policy.
Sources #
Primary source and supporting public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.