Xi concluded his AI speech by quoting the "Discourses on Salt and Iron," an 81 BC text about the Han dynasty's struggle to govern the privatized salt and iron industries. This historical parallel suggests China's current AI governance debate is a modern iteration of a centuries-old tension between private innovation and state extraction/control.
A US-style "Project Glasswing" for AI safety is unlikely in China because its major tech firms (Alibaba, Tencent, ByteDance) are fierce rivals. Unlike in the US, where a lab could partner with a company like Microsoft, Chinese conglomerates would refuse to hand over proprietary model technology to direct competitors, whom they actively poach from and have long-standing rivalries with.
Xi Jinping's emphasis on "openness" in AI may be primarily aimed at a domestic audience. It serves to reassure Chinese AI researchers and developers that a harsh government crackdown, like the one against Jack Ma, is not imminent. This signaling encourages innovation by making them feel the government is on their side, at least for now.
China's seemingly contradictory AI policy—promoting both openness and strict control—stems from its view of AI as a "normal technology" akin to the internet. Unlike some in the West who see it as an existential risk, Beijing sees a technology that should be diffused as widely as possible for economic gain, while managing its risks with conventional tools like robust cybersecurity.
China's strategy of releasing powerful open-weight models to "fast follow" US capabilities creates a paradox. While it closes the technology gap, it prevents Chinese labs from building sustainable businesses. Without monetizing via proprietary APIs like OpenAI, they cannot generate the revenue needed to acquire the massive compute resources required for long-term competition.
The Chinese government may be more willing to release powerful open-weight AI models because it already operates in an environment of "cyber exploit abundance." Having long managed a system of acquiring and using zero-day vulnerabilities, the state may not perceive the same level of novel threat from democratized cyber capabilities as Western governments do.
Analysts should be skeptical of Xi's promises of global AI cooperation, drawing parallels to his 2017 Davos speech. At Davos, he championed free trade just before imposing harsh economic coercion on South Korea. This history suggests his lofty rhetoric on AI may serve immediate diplomatic goals while masking more confrontational long-term strategies.
Xi's speech targeted the Global South with promises of "inclusiveness." However, China's new, more powerful models like Kimi are significantly more expensive to run than previous generations. This creates a tension: while China offers a political alternative to US AI dominance, the rising costs of its frontier technology may make it unaffordable for the very developing nations it courts.