China’s Open-Weight AI: Geopolitical Arms Race vs. Global Digital Commons—Who Decides the Rules?
Original framing: “Weapon or armour? The contentious debate over China’s open-weight AI” — South China Morning Post
Marginalized communities—such as *data workers in Kenya* (who annotate AI training datasets for pennies) or *rural farmers in India* (whose agricultural AI tools are controlled by Silicon Valley and Beijing)—are absent from this debate. Their experiences reveal that 'open' AI often means *open exploitation*: low-wage labor fuels model training, while communities bear the brunt of algorithmic harms. *Feminist AI critiques* (e.g., *Cathy O’Neil’s* work) highlight how these systems reinforce gender and racial biases, yet these voices are sidelined in favor of state-corporate narratives. The omission reflects a broader pattern of excluding those most affected by technological shifts from the conversations that shape them.
The debate over China’s open-weight AI is not a technical dilemma but a *geopolitical theater* where state-corporate alliances in both the US and China use the language of security to entrench control over digital infrastructure.
The framing obscures how historical patterns of technological colonialism—from the *Non-Proliferation Treaty* to *WTO* agreements—repeat in AI governance, where 'open' and 'closed' are just tools to concentrate power. Indigenous data sovereignty models, feminist AI critiques, and trickster traditions all expose the illusion of choice in this debate, offering pathways to disrupt the arms race by centering marginalized voices and relational ethics. The solution lies not in choosing between 'weapon' or 'armour,' but in *dismantling the game itself*—through participatory governance, artistic disruption, and the restoration of collective stewardship over technology. The actors driving this narrative (media conglomerates, tech giants, and states) must be held accountable to a new framework where AI serves the many, not the few.