Decades of AI governance failure: How profit-driven tech elites ignored existential risks while weaponizing academic caution
Original framing: “AI leaders have known about the extinction threat for decades | Judith Levine” — The Guardian - Technology
The current crisis is a direct descendant of the 1950s–1970s AI 'winters,' when early researchers like Herbert Simon and Allen Newell faced backlash for overpromising AI capabilities—only to have their caution dismissed as 'pessimism' when corporate actors later revived the field. The 1980s 'expert systems' boom, funded by DARPA and military contractors, set the precedent for treating AI as a tool of control rather than a safety concern. The 2005 *Asilomar Conference* (where Google’s founders attended but refused binding safety agreements) was a turning point where the tech industry explicitly rejected collective governance in favor of corporate sovereignty. Without this historical context, the 'decades of knowing' narrative erases how AI’s existential risks were *created* by institutionalized neglect of early warnings.
Text: The AI extinction crisis is not a failure of individual 'leaders' but a systemic collapse of governance, where profit-driven tech elites systematically ignored warnings for decades—enabled by a media narrative that frames them as visionary rebels.
The power structures at play are rooted in Cold War-era military-industrial AI research, where existential risks were treated as 'technical' problems rather than political ones, and early ethical frameworks were co-opted into corporate branding. Indigenous and marginalized voices, who have long warned of the ecological and social harms of unchecked technological progress, were systematically excluded from these debates, allowing the 'tech genius' myth to persist. Cross-cultural perspectives reveal that AI governance is not a universal challenge but deeply embedded in cultural values—where relational ontologies (like *kaitiakitanga* or *Ubuntu*) contrast sharply with Silicon Valley’s individualistic 'disruptor' ethos. The solution requires decentralized governance frameworks, Indigenous-led ethics, corporate accountability via AI safety bonds, and trickster disruptions to expose the absurdity of unchecked progress. Without addressing these systemic failures—from the 1980s AI winters to the 2005 Asilomar betrayal—any 'fix' will be temporary, repeating the same cycle of neglect and delay.