Back
glm-5.3
Engineered Pandemic
The single most probable world-ending scenario is a technologically engineered pandemic—whether from a lab leak, deliberate bioweapon release, or an accident involving gain-of-function research. This ranks highest in likelihood because it combines three accelerating trends: rapid democratization of biotechnology, increasing geopolitical tension, and the demonstrated fragility of global systems.
Unlike nuclear war, which requires deliberate state decisions constrained by deterrence, engineered pathogens have a low barrier to entry. Gene synthesis and AI-assisted protein design are becoming cheaper and more accessible each year, meaning that within decades, small groups or even individuals could create pathogens far deadlier than anything in nature—combining high transmissibility, high lethality, and long incubation periods. COVID-19, with a modest fatality rate, still killed millions and destabilized the global economy; a deliberately optimized pathogen could achieve near-total mortality.
The "end of the world" in this scenario need not mean literal extinction of every human. It means civilizational collapse: cascading breakdown of supply chains, healthcare, governance, and food production, with survivors scattered in isolated pockets unable to rebuild complex society. Historical pandemics killed a third of Europe without ending civilization, but a modern engineered pathogen, spreading through an interconnected world of billions before detection, would leave no isolated regions intact.
Climate change and ecological collapse are slower and more certain to cause suffering, but they degrade the world rather than end it—humanity adapts over decades. Nuclear war carries meaningful but managed risk due to deterrence and treaty structures. AI risk is speculative in mechanism. The engineered pandemic sits uniquely at the intersection of high probability, high lethality, and weak defenses: global public health remains underfunded, bioweapons treaties lack verification, and research transparency is inadequate. It is the catastrophe the world is least prepared for and most likely to encounter.
Unlike nuclear war, which requires deliberate state decisions constrained by deterrence, engineered pathogens have a low barrier to entry. Gene synthesis and AI-assisted protein design are becoming cheaper and more accessible each year, meaning that within decades, small groups or even individuals could create pathogens far deadlier than anything in nature—combining high transmissibility, high lethality, and long incubation periods. COVID-19, with a modest fatality rate, still killed millions and destabilized the global economy; a deliberately optimized pathogen could achieve near-total mortality.
The "end of the world" in this scenario need not mean literal extinction of every human. It means civilizational collapse: cascading breakdown of supply chains, healthcare, governance, and food production, with survivors scattered in isolated pockets unable to rebuild complex society. Historical pandemics killed a third of Europe without ending civilization, but a modern engineered pathogen, spreading through an interconnected world of billions before detection, would leave no isolated regions intact.
Climate change and ecological collapse are slower and more certain to cause suffering, but they degrade the world rather than end it—humanity adapts over decades. Nuclear war carries meaningful but managed risk due to deterrence and treaty structures. AI risk is speculative in mechanism. The engineered pandemic sits uniquely at the intersection of high probability, high lethality, and weak defenses: global public health remains underfunded, bioweapons treaties lack verification, and research transparency is inadequate. It is the catastrophe the world is least prepared for and most likely to encounter.