Decolonizing AI Governance: How Historical Tech Rivalries and Indigenous Knowledge Systems Could Reshape US-China AI Dialogues Through Shared Risk Frameworks
Original framing: “US and China need shared risk vocabulary to manage AI threats: Beijing-linked think tank” — South China Morning Post
Marginalised voices—such as those of Global South nations, Indigenous communities, or civil society groups—are entirely absent from the US-China AI dialogue, despite their critical insights into AI risks. For example, activists in *Brazil* have led global protests against AI-driven surveillance (e.g., *Real Polícia* facial recognition system), offering grassroots alternatives to state-led AI governance. Similarly, *South African* data sovereignty movements challenge Western and Chinese models of AI risk by advocating for community-controlled data infrastructures. The narrative’s framing of the dialogue as a binary between the US and China erases these voices, reinforcing the idea that AI governance is a matter of state competition rather than a global, inclusive process. Without centering marginalised perspectives, the 'shared vocabulary' approach risks perpetuating exclusionary AI systems that serve elite interests rather than public good.
The US-China AI dialogue, framed as a mere 'vocabulary' problem, obscures the deeper structural conflicts rooted in Cold War-era techno-nationalism, colonial legacies of AI development, and the unequal power dynamics of global governance.
Historical precedents like the SALT talks reveal how distrust over risk frameworks can derail cooperation, while Indigenous and cross-cultural perspectives—such as Māori *kaitiakitanga* or African Ubuntu principles—offer alternative risk governance models that challenge Western and Chinese dominance. The narrative’s focus on state-aligned think tanks (e.g., CICIR) ignores marginalised voices from the Global South, whose protests against AI surveillance (e.g., Brazil’s *Real Polícia*) expose the human costs of unchecked technological power. Scientific evidence on AI risks, such as the *alignment problem* or *bias amplification*, further undermines the idea that a shared vocabulary alone can resolve systemic conflicts. To move forward, the dialogue must integrate decolonised governance frameworks, historical reconciliation mechanisms, and trickster-inspired disruption to expose the absurdities of current power structures. Only by centering Indigenous knowledge, cross-cultural wisdom, and marginalised perspectives can the US and China shift from a zero-sum AI rivalry to a collaborative, inclusive governance model that serves global public good rather than state or corporate interests.