U.S., other nations back open-source AI with 'strong security' at China summit

• Open-source AI models from Chinese companies like DeepSeek and GLM 5.2 are free to download and use, contrasting with closed, pay-to-use models from U.S. firms like Anthropic
• The statement signals Asia-Pacific's move away from closed models toward open-weight systems combined with state-coordinated energy, telecom, and digital infrastructure
• The agreement bridges 'open vs. closed' debate by focusing on building trusted open ecosystems with testing, transparency, data protection, and deployment controls acceptable to governments and enterprises

Summary: APEC Nations Support Open-Source AI with Security Requirements

Key Development:

The 21 APEC member economies, including the U.S. and China, issued a joint statement backing open-source AI development with "strong security assurance through development and deployment." The declaration emphasizes security, data protection, and intellectual property rights—marking the first ministerial-level agreement on open-source AI cooperation.

Main Players:

• China: Leading recent open-source AI development through companies like DeepSeek and GLM 5.2, which offer free, downloadable models
• U.S. companies: Primarily offering closed, paid AI models (e.g., Anthropic)
• Chinese Minister Li Lecheng chaired the Thursday meeting in Chengdu

Market Implications:

The statement signals a significant shift in open-source AI from libertarian roots toward increased state oversight. Winston Ma, NYU adjunct law professor, noted Asia-Pacific is moving from closed models to "open-weight systems" combined with state-coordinated infrastructure in energy, telecom, and digital sectors.

Analysts view the agreement as meaningful given U.S.-China tensions. Wei Sun of Counterpoint Research highlighted that "strong security assurance" language allows security-conscious governments to support open models while maintaining testing, transparency, and deployment controls.

Strategic Context:

The framework addresses growing government desires to control AI capabilities as both Washington and Beijing reportedly seek tighter controls on domestic AI development. The focus has shifted from "open vs. closed" debates to building trusted open ecosystems acceptable for government and enterprise deployment.

Future Consideration:

APEC's Business Advisory Council urged ministers to prepare for quantum computing's impact on encryption and financial system security.