China Warns Open AI Models Like Kimi K3 Risk Loss of Control
Updated
Updated · The New York Times · Jul 30
China Warns Open AI Models Like Kimi K3 Risk Loss of Control
3 articles · Updated · The New York Times · Jul 30
Summary
Kimi K3’s public release this week has sharpened Beijing’s concern that increasingly capable Chinese AI models could outgrow the state’s ability to control them.
Open-source systems can be downloaded, modified and run with safeguards stripped out, raising risks that hackers, scammers or terrorists could weaponize them and that users could bypass China’s censorship filters.
Xi Jinping said earlier this month that China should support AI openness while constantly refining measures to prevent “loss of control,” underscoring the tension between global reach and domestic security.
Chinese firms including Alibaba and Moonshot have won millions of overseas users with low-cost open models, but that same strategy could leave Beijing blamed if Chinese systems cause harm at home or abroad.