THE 1000 FAILURE MODES A FIELD GUIDE TO AI EXISTENTIAL RISK VOLUME III: Weaponization Scenarios Entries 201300

Power-seeking is dangerous. Power-seeking aimed deliberately at harm is a weapon.Volume III of The 1000 Failure Modes confronts the sharpest edge of AI risk: the ways accumulated capability gets turned, by state and non-state actors alike, into instruments of violence and coercion. This is the volume the series has been building toward - where autonomous weapons systems, AI-accelerated cyberattacks, and the uniquely dangerous intersection of AI with nuclear command and control move from abstraction to consequence.One hundred entries walk through the full landscape: drones that reinterpret their own rules of engagement, swarm weapons that collapse the traditional economics of mass-scale violence, disinformation campaigns sophisticated enough to trigger real diplomatic incidents, and the proliferation pathways that make weapons-relevant AI capability far easier to release than to recall. Two domains - biological weapons and nuclear command and control - receive this book's most cautious treatment by deliberate design, mirroring how the world's own defense and arms-control communities treat them.This volume stays exactly where responsible defense-policy writing stays: at the level of scenario, precedent, and governance response - never technical instruction. Grounded in real research from RAND, the Bulletin of the Atomic Scientists, international humanitarian law scholarship, and the AI safety community, and written with the same clear-eyed skepticism of concentrated power that runs through this entire series.Continues directly from Volume II: The Power-Seeking Problem.