21 APEC Economies Endorse Open-Source AI with 'Strong Security Assurance' at Chengdu Summit

The 21 APEC member economies—including the United States and China—released a 'Chengdu statement' on AI at a digital ministers' summit in China on July 23, endorsing support for open-source AI development while emphasizing 'security, data protection and intellectual property rights.' Li Lecheng, China's Minister of Industry and Information Technology, noted this is the first multilateral APEC AI statement to explicitly include open-source cooperation at the minister level.

The statement reflects a subtle but significant governance shift: both Washington and Beijing are seeking tighter domestic controls on frontier-model development, yet publicly endorsing open-weight models as trustworthy when built with 'strong security assurance.' The phrase gives security-conscious economies diplomatic cover to support open models while still demanding testing, transparency, data protection and deployment oversight.

For practitioners, this signals tacit agreement between rivals that open-source AI, when vetted, can coexist with responsible AI governance. The APEC consensus also underscores that neither superpower can unilaterally dictate global AI norms—the open-weight surge from Chinese labs this year (Moonshot, DeepSeek) has already shifted the table.