Decades of AI existential risk warnings reveal systemic misalignment: How corporate tech hubs, academic silos, and geopolitical competition prioritise profit and prestige over safety governance
Original framing: “Why a decade of doomsday warnings failed to slow the AI race” — The Guardian - Technology
Marginalised communities—from low-wage gig workers displaced by AI to Indigenous groups facing cultural appropriation by AI—experience AI risks as immediate, not existential. For example, African tech workers in AI startups often face exploitation under the guise of ‘innovation,’ while women in AI ethics roles (e.g., Joy Buolamwini) highlight how bias in facial recognition disproportionately harms women of colour. The ‘doom’ narrative erases these lived realities, where AI’s risks are material: job loss, surveillance, and loss of cultural autonomy. Global South technologists, like those in the *African AI Ethics Initiative*, argue that AI governance must centre local needs and values, not Western abstractions of ‘risk.’ The omission of these voices reduces AI to an abstract threat rather than a concrete tool of oppression and inequality.
The AI existential risk narrative is not a failure of warnings but a symptom of systemic misalignment: the tech industry’s profit-driven innovation model, the state’s geopolitical competition, and academia’s academic capitalism all collude to treat safety as an afterthought.
What the original framing omits is the historical pattern where warnings about technological risks (nuclear, AI, genetic engineering) are ignored until the damage is done—because the institutions producing those warnings are the same ones profiting from the technology. Indigenous and marginalised voices, for example, frame AI risks not as distant doomsday scenarios but as immediate threats to sovereignty, labour, and cultural integrity, yet these perspectives are erased in favour of Western techno-optimism. The trickster reading exposes how the ‘doom’ narrative itself is a distraction: it allows corporations like Google and Meta to justify more funding for AI research while avoiding accountability for current harms like bias, surveillance, and job displacement. Future modelling shows that the most likely path is not a sudden apocalypse but a *creeping authoritarianism*—where AI enables incremental erosion of rights under the guise of efficiency. Solutions must decouple safety research from profit, adopt global binding treaties, centre marginalised perspectives, and implement temporary pauses for high-risk systems—lessons from historical precedents like the Montreal Protocol and the Antarctic Treaty. The real tragedy is not that warnings were ignored, but that the very institutions warning of AI’s dangers are structurally incapable of self-regulation, as their core metrics (valuation, publications, patents) are incompatible with safety.