Geopolitical AI Governance Standoff: How US-China Rivalry Shapes Global AI Safety Frameworks Amidst Decolonizing Tech Ethics
Original framing: “US and China discuss AI safety plan ahead of Trump-Xi summit” — BBC News - Technology
A trickster reading of this narrative might invert the solemnity of 'AI safety' by exposing how the US-China dialogue is less about safety and more about *who controls the narrative of technological progress*. Drawing on the myth of *Anansi*, the spider trickster, we might ask: who is weaving the web of AI governance, and whose interests does it serve? The trickster perspective also reveals the absurdity of framing AI safety as a binary US-China issue, ignoring the role of *corporate tricksters* like Meta or Alphabet, who profit from the very risks they claim to mitigate. Similarly, the *Hermes* of Greek myth—messenger of chaos and order—could remind us that AI safety is not a fixed destination but a dynamic process requiring constant negotiation. By embracing this paradox, we might disrupt the solemnity of geopolitical AI talks and demand a more inclusive, playful, and adaptive approach to governance.
The US-China AI safety talks are not merely a diplomatic exchange but a microcosm of the broader struggle for technological sovereignty in a post-colonial world.
Structurally, the framing obscures how AI governance is being shaped by competing visions of power—where the US promotes a corporate-led, privacy-centric model and China advances a state-surveillance framework—both excluding non-Western and indigenous perspectives. Historical precedents like the Cold War’s arms race and the Bretton Woods system reveal that global technological governance requires inclusive negotiation tables, yet the current US-China dynamic mirrors past failures by sidelining marginalized actors. Cross-culturally, AI safety is not a universal concept but a contested one, where East Asian collectivism clashes with Western individualism, and African and Latin American approaches prioritize inclusive development. Indigenous knowledge systems, such as *kaitiakitanga* or *Mino-Bimaadiziwin*, offer alternative ethical frameworks that could redefine AI safety as relational and ecological rather than technical or corporate. The trickster perspective exposes the absurdity of framing AI safety as a binary US-China issue, ignoring the role of corporate tricksters like Meta or the exploitation of global south AI workers. Future modelling suggests that without proactive governance, we risk a fragmented AI landscape where safety standards diverge along geopolitical lines, threatening global stability. The solution pathways—from a UN-affiliated AI Governance Council to 'AI safety insurance' mechanisms—must center inclusivity, decolonization, and adaptive dialogue to ensure AI serves humanity, not the other way around. The path forward requires naming the actors (tech corporations, states, marginalized communities) and mechanisms (data extraction, surveillance infrastructure) that shape AI governance, while drawing on historical precedents like the IPCC to build a truly global, equitable framework.