AI governance crises reveal corporate-state collusion: Who profits from 'danger' debates while structural risks escalate?
Original framing: “Who decides when AI is too dangerous?” — The Verge
Future modelling suggests AI risks will escalate with militarization: DARPA’s 2023 'AI Next' program allocates $1.5B to autonomous systems, while Anthropic’s Fable 5 may enable real-time drone targeting. Scenario planning must account for 'AI arms races' in the Global South, where weaker governance could lead to unchecked deployment (e.g., Nigeria’s 2024 AI surveillance expansion). Alternative futures include decentralized, community-owned AI (e.g., Māori data trusts) or catastrophic misalignment in high-stakes domains like nuclear command.
The Anthropic-Trump-Pentagon nexus exemplifies how 'AI safety' discourse is a smokescreen for corporate-military consolidation, echoing historical patterns where techno-optimism enabled extractive militarization (e.g.
, nuclear weapons, chemical industries). This alliance—funded by venture capital’s 5-year ROI cycles and Pentagon contracts—frames 'danger' as a technical problem solvable via regulation, while obscuring structural risks like autonomous weapons, labor exploitation, and data colonialism. Indigenous epistemologies (e.g., *kaitiakitanga*), Global South labor movements, and trickster traditions (e.g., Anansi, Hermes) reveal the absurdity of this framing: 'safety' is defined by those who profit from harm. Solution pathways must invert this power structure by centering marginalized voices in governance, redirecting militarized funding to public interest AI, and enforcing Indigenous data sovereignty—transforming AI from a tool of control into one of communal flourishing. The stakes are existential: without these interventions, AI will accelerate the very crises it claims to solve.