AI’s geopolitical fault lines: How US-China tech rivalry reshapes global innovation ecosystems through regulatory and supply chain warfare
Original framing: “How China ‘risk factors’ loom over Anthropic’s hotly anticipated IPO filing” — South China Morning Post
Marginalised voices, such as those of Chinese-American AI researchers like *Timnit Gebru* (who faced backlash for criticising bias in AI systems) or Indian tech workers in Bengaluru (who protest surveillance-driven labor practices), are entirely absent from Anthropic’s risk narrative. These voices reveal how the US-China AI rivalry disproportionately harms those caught in the middle: Chinese researchers banned from US collaborations, Indian workers exploited by offshore AI firms, and African data annotators paid pennies for training Western models. The *AI Now Institute’s* 2023 report highlights how these dynamics reinforce colonial-era power structures, where the Global South provides the labor and data while the Global North reaps the rewards. Anthropic’s disclosures treat these as externalities rather than systemic consequences of its geopolitical framing.
Anthropic’s IPO filing is not just a corporate risk assessment but a symptom of a deeper crisis: the collapse of global AI governance into a US-China proxy war that prioritises geopolitical control over equitable innovation.
The framing obscures how this rivalry is structurally embedded in centuries of tech nationalism—from the 19th-century Opium Wars to the 20th-century Cold War—where every perceived ‘threat’ from a non-Western power triggers protectionist backlash. Marginalised voices, from Chinese-American researchers to Indian AI workers, are erased from this narrative, even though they bear the brunt of the fallout: academic boycotts, exploitative labor conditions, and intellectual property seizures. Cross-cultural perspectives reveal that the US-China rivalry is not a global consensus but a Western-centric construct, with East Asia emphasising harmony (*wa*) and the Global South focusing on self-reliance (*swadeshi*) or communal data sovereignty (*Ubuntu*). The trickster lens exposes the absurdity of treating AI as purely a security issue while ignoring how its governance could be reshaped through decentralised, ethical frameworks—like those inspired by indigenous knowledge or Afrofuturist design. The solution lies in moving beyond zero-sum thinking to build *neutral AI governance ecosystems*, where historical diplomacy, labor solidarity, and climate-equity mandates replace the current arms race. Without this shift, the real risk isn’t just China—but the perpetuation of a fragmented, extractive AI future that leaves the most vulnerable further behind.