Abstract China is not a liberal democracy with independent corporations. It is a state where strategically significant technology companies do not stand apart from state will, not legally, not institutionally, not historically. It is a state that treats technological strength as a tool of global influence, a direct competi- tor to the Western order, widening its reach across technology, the military, and the economy. Against that background sits a question Western institutions prefer not to ask out loud. Why does pre- cisely this kind of state push open AI models, free or at dumping prices, onto world markets so hard? The comfortable answer is that independent researchers want to advance science and open up competition. It is an appealing story. It also assumes that Chinese AI companies have the same independence as a startup in Berlin or San Francisco. They do not. Open weights joined to closed training data are not transparency. They are the look of it, and the look is the way in. Cheap tokens do the rest: they move Western business and government onto these models and build dependence before anything is ever switched on. This is the Economic Deployment Vector, price as a path to infrastructure capture rather than a market tactic. The capability behind it is not hypothetical. Peer-reviewed work shows that a backdoor can be planted in training through a tiny number of documents, and that it survives later safety training. It can then lie dormant for years and be triggered on command, because the systems that adopt these models stay online. It can also hide where Western auditing does not look, in the Chinese-weighted part of a bilingual model that English-only evaluation never reaches. This is the Bilingual Blind Spot. And a clean record across versions is no reassurance, because earning trust before striking is the known shape of this kind of attack. The conclusion follows. Reliable detection may be impossible, or so costly it eats the saving that made the model attractive, so the answer is not better auditing but less dependence: your own models, or a trusted state's, never one whose institutions leave no room to refuse. This is sharp toward China, and fair, because it rests on the structure of the state, not on origin. Two of the three technical links are already established. The third, that the same cheap method reaches sophisticated behaviour, is neither proven nor refuted, and the burden of showing it cannot be done falls on whoever would relax because of it.
Tigran Margaryan (Mon,) studied this question.