“The Precipice: Existential Risk and the Future of Humanity” is a 2020 book by the Oxford philosopher Toby Ord, then a senior research fellow at the Future of Humanity Institute. It is the most prominent book-length treatment of existential risk - threats that could permanently destroy or drastically curtail humanity’s long-term potential - and it brought the topic to a wide general audience. Ord argues that humanity has entered a uniquely dangerous period, which he calls the Precipice, beginning roughly with the first atomic bomb test in 1945, in which our destructive power has outrun our wisdom and restraint.
The book’s most-discussed figure is Ord’s estimate of a 1-in-6 chance of an existential catastrophe over the next hundred years. Strikingly, he attributes the single largest share of that risk - around 1 in 10 over the century - to unaligned artificial general intelligence, which he places above engineered pandemics, nuclear war, and climate change. The estimates are deliberately rough and meant to direct attention and effort, not to be read as precise predictions; Ord’s broader point is that the resources society devotes to reducing these risks are tiny relative to what is at stake.
The Precipice is closely tied to longtermism, the ethical view that positively shaping the long-term future is among the most important things we can do, given how many future generations could exist. The book helped move AI safety from a niche concern into mainstream policy and philanthropic discussion, and it is frequently cited - and contested - in debates with those who, like the “AI as normal technology” camp, regard such catastrophe estimates as overstated.
Why a general reader should care: the framing and risk estimates in this book underpin much of the modern argument for treating AI safety as a civilizational priority, an argument now shaping real funding and regulatory decisions.