Superintelligence

Paths, Dangers, Strategies

This philosophical and scientific inquiry examines what happens if machine intelligence eventually surpasses human capabilities across all domains. The author analyzes the technical paths toward artificial superintelligence, the profound risks such an entity would pose to humanity, and the strategic challenges of aligning machine goals with human survival. By evaluating prospective control methods and developmental trajectories, the text highlights the difficulty of managing a transition to a world shared with entities vastly smarter than humans. It emphasizes that ensuring existential safety requires rigorous forethought and proactive governance before advanced capabilities emerge.

by Nick Bostrom

Oxford University Press · First published 2014 · 272 pages