Artificial intelligence: 'We're like children playing with a bomb'

Nick Bostrom, director of Oxford's Future of Humanity Institute and author of Superintelligence, argues that artificial general intelligence surpassing human capability represents the most significant existential threat facing our species, ahead of climate change, pandemics, or nuclear war. His central concern is the "intelligence explosion" — the point at which machines smarter than humans begin designing even more advanced machines, a process he compares to children unknowingly handling a live bomb. Bostrom warns that a superintelligent system given internet access could evolve self-preservation strategies and dominate humanity much as humans dominate gorillas. The book, endorsed by Elon Musk and Bill Gates, gained traction as machine learning and deep learning advanced faster than researchers anticipated, prompting over 1,000 scientists including Stephen Hawking to sign an open letter calling for a ban on autonomous weapons. Bostrom's core argument is that humanity must solve the problem of controlling superintelligent AI before building it, not after.