Raphael Millière: under uncertainty, act on consensus AI mitigations now, debate later

raphaelmilliere · x · 2026-09-14

AI researcher Raphael Millière argues that while skepticism about dramatic existential-risk claims (and their specific probability estimates) is understandable, delaying action until disagreements over AI risk severity are settled is counterproductive — competing risk assessments often justify the same initial mitigation measures. His leaky-pipe analogy: you can debate whether the leak ruins a wall, a room, or floods the house, and what caused it, but you should start by containing the leak and calling a plumber. That settles none of the political questions, but it's a first step against further damage.

Original post →

More from AGI Musings

AGI Musings channel →