What happens if artificial intelligence becomes more capable than the institutions-and people-responsible for controlling it?
The Ten-Year Corridor is an accessible, evidence-led examination of how advanced AI could develop over the next decade and the decisions that may determine its impact. Written for both technical and general readers, it explains the alignment problem, the limits of current training methods, and the difference between a powerful tool and a system able to pursue objectives with increasing independence.
The book follows a plausible shared timeline before branching into several possible futures: successful international restraint, a treaty weakened by political exceptions, growing dependence on AI-managed infrastructure, rivalry between competing systems, a false alarm in which the danger was overstated, and a catastrophic loss of human control.
Rather than presenting one prediction as inevitable, Michael Yoannou separates established facts from uncertainty and speculation. Each scenario asks what governments, laboratories, researchers, and the public might do-and what would have to go wrong, or right, for that future to emerge.
The Ten-Year Corridor is a book about artificial intelligence, but also about incentives, power, coordination, and the narrow period in which human choices may still matter most.