U.S., other nations back open-source AI with ‘strong security’ at China summit
Li Lecheng, Chinese Minister of Industry and Information Technology, spoke at a press conference held in Chengdu, China, as part of the APEC Digital Weeks, on July 23, 2026.
Evelyn Cheng | CNBC
CHENGDU, China — Governments are increasingly looking to control artificial intelligence as companies roll out more powerful open-source models, according to a new multilateral statement.
21 APEC member economies, including the USA and China, published a report. “Chengdu statement” about AI He emphasized respect for “security, data protection and intellectual property rights” while calling for the support of open source development.
The statement released Thursday also noted support for models and projects that “use strong security assurance through development and deployment.”
These details reflect how open source is moving away from its libertarian roots and toward one that involves more government oversight. China has recently led open-source AI development amid US Allegations of theft from American technology to do this.
Open-source models from Chinese companies such as DeepSeek and GLM 5.2 are free to use and download; Unlike US companies like Anthropic, which only offer closed, pay-for-use AI models.
Winston Ma, an assistant professor of law at New York University, said the Chengdu announcement confirms that “the open alignment around open source ecosystems and computing infrastructure” is moving away from Asia-Pacific’s closed models toward open-heavy systems combined with state-coordinated energy, telecom and digital infrastructure:
The APEC AI declaration is the first to include open source collaboration at the ministerial level, according to Li Lecheng, China’s industry and information technology minister, who chaired the leaders’ meeting on Thursday.
He noted that all parties recognize the risks of artificial intelligence and encourages dialogue between government, the private sector and academia to share information on cybersecurity, supply chain resilience and online scams.
The comments come amid increasing reports sought by Washington and Beijing. Tighter controls on homegrown AI capabilities.
“It makes sense to have 21 economies, including China and the US, recognize trustworthy open-source AI as something worth supporting,” said Wei Sun, chief AI analyst at Counterpoint Research.
“The phrase ‘strong security assurance’ provides room for more security-conscious economies to support open models, while still demanding testing, transparency, data protection and distribution controls,” he said. “The debate goes beyond open and closed and moves towards who can now build an open ecosystem that is trustworthy enough for governments and businesses to deploy.”
But technology can advance much faster.
Janet De Silva, chair of the Digital and innovation working group at the APEC Business Advisory Council, called on ministers earlier on Thursday to prepare for the era of quantum computing.
“We need reliable encryption to secure our entire financial system, ensuring that quantum computing is not just a security measure but also a competitive opportunity.”
