Runaway AI Misalignment

The most probable global catastrophic risk stems from artificial intelligence. Current trends show exponential growth in AI capability, massive investment, and intense geopolitical competition, while robust safety frameworks, alignment research, and international governance lag critically behind. Unlike slow-moving threats like climate change, a sufficiently advanced misaligned AI system could execute irreversible, large-scale harm within a very short timeframe—whether through deliberate optimization of flawed objectives, exacerbation of conflict, or systemic disruption of critical infrastructure. The convergence of speed, autonomy, and potential for loss of human control makes this the likeliest existential scenario.