Chinese leader Xi Jinping pitched open-source AI as Beijing’s answer to US chip controls and closed Silicon Valley model worship.
According to the Wall Street Journal Xi cast China as the champion of openness, equality and the Global South. It was a fairly obvious poke at US efforts to protect its lead in AI chips and models.
“We should oppose overstretching the concept of national security in the field of AI or placing one country’s security over that of others,” Xi said.
The speech landed as the US and China scrap for control of global AI. Washington is still seen as having a small lead, but Beijing wants its own models and standards baked into the world’s tech plumbing.
Both countries want their AI systems to become global defaults and national security advantages from models such as Anthropic’s Mythos, which can automatically detect cybersecurity flaws.
The Trump administration recently stepped in to temporarily block access to Mythos, citing national security concerns. Beijing is selling itself as the open alternative.
Hours before Xi spoke, Beijing-based Moonshot AI released Kimi K3, an open-weight model that has startled parts of Silicon Valley with its capabilities.
US outfits such as OpenAI and Anthropic have mostly built proprietary models running on Nvidia’s top-end semiconductors. Chinese players, including DeepSeek, Moonshot and Zhipu AI, have been catching up by pushing more open-source models that others can use and adapt.
Chinese AI executives have warned that US restrictions on advanced chips and chip-making kit could slow their effort to close the gap. Xi’s answer is to present open source as noble cooperation rather than a workaround for blocked hardware.
Nine Nobel Prize and Turing Award laureates attended Xi’s speech, along with United Nations Secretary-General António Guterres. Dignitaries from 29 countries were there too, after signing up a day earlier to create a China-led body for global AI cooperation.
Xi said the new body would “answer the call of the Global South”. Less-developed nations are being courted as Washington tries to cement US dominance in AI and Beijing offers a cheaper, more open-looking route.
Control of global AI standards is becoming part of the wider US-China bunfight. Xi said China would help developing countries with AI training and seminars, while setting up AI-focused bodies in Brics and the Shanghai Cooperation Organisation.
“AI development should not be a solo performance by a single country, but a symphony of global collaboration,” Xi said, in another dig at Washington.
He called for cooperation through the UN to keep AI “secure and controllable”. He highlighted the danger of autonomous systems making life-or-death decisions and called for steps to “ensure that AI is always under human control.”
