The Economist wrote an article on existential Risk. You can read it here:
The most relevant extract:
In a recent book, “The Precipice”, Toby Ord of the Future of Humanity Institute at Oxford University defines an existential risk as one that “threatens the destruction of humanity’s long-term potential”. Some natural disasters could qualify. An impact like that of the 10km asteroid which ushered out the dinosaurs 66m years ago is one example. A burst of planet-scouring gamma rays from a nearby “hypernova” might be another. A volcanic “super-eruption” like the one at Yellowstone which covered half the continental United States with ash 630,000 years ago would probably not extinguish the human race; but it could easily bring civilisation to an end. Happily, though, such events are very rare. The very fact that humans have made it through hundreds of thousands of years of history and prehistory with their long-term potential intact argues that natural events which would end it all are not common.
Do you feel lucky, punk?
For already existing technologically mediated risks, such as those of nuclear war and climate collapse, there is no such reassuring record to point to, and Mr Ord duly rates them as having a higher chance of rising to the existential threat level. Higher still, he thinks, is the risk from technologies yet to come: advanced bioweapons which, unlike the opportunistic products of natural selection, are designed to be as devastating as possible; or artificial intelligences which, intentionally or incidentally, change the world in ways fundamentally inimical to their creators’ interests.
No one can calculate such risks, but it would be foolish to set them at exactly zero. Mr Ord reckons almost anyone looking at the century to come would have to concede “at least a one in 1,000 risk” of something like a runaway ai either completely eradicating humanity or permanently crippling its potential. His carefully reasoned, if clearly contestable, best guesses lead him to conclude that, taking all the risks he cites into account, the chances of humankind losing its future through such misadventure in the next 100 years stands at one in six. The roll of a single die; one spin of the revolver chamber.
Mr Ord is part of a movement which takes such gambles seriously in part because it sees the stakes as phenomenally high. Academics who worry about existential risk—the trend began, in its modern form, when Nick Bostrom, a Swedish philosopher, founded the Future of Humanity Institute in 2005—frequently apply a timeagnostic version of utilitarianism which sees “humanity’s long-term potential” as something far grander than the lives of the billions on Earth today: trillions and trillions of happy lives of equal worth lived over countless millennia to come. By this logic actions which go even a minuscule way towards safeguarding that potential are precious beyond price. Mr Ord, one of the founders of the “effective altruism” movement, which advocates behaviour rooted in strongly evidence-based utilitarianism, sees a concern with existential risk as part of the same project.
Risks that are merely catastrophic, not existential, do not tend to be the subject of such philosophical rumination. They are more amenable to the sort of calculations found in the practice of politics and power. Take the risk of a nuclear attack. According to Ron Suskind, a reporter, in November 2001 Dick Cheney noted that America needed new ways to confront “low-proba-bility, high-impact” events. “If there’s a 1% chance that Pakistani scientists are helping al-Qaeda build or develop a nuclear weapon,” the vice-president said, “we have to treat it as a certainty in terms of our response”. Such responses included new wars, new government agencies (the Department of Homeland Security) and new executive powers, including warrantless surveillance.
If every perceived 1% risk were met with such vigour the world would be a very different place—and not necessarily a safer one. But it is striking that some risks of similar magnitude are barely thought about at all. Imagine Mr Cheney was considering the possibility of losing a city over a 20-year period. What else puts a city’s worth of people and capital at a 1% risk every few decades? The sort of catastrophic risks that crop up every millennium or so, threatening millions of lives and cost trillions. Perhaps they should be treated equally seriously. As Rumtin Sepasspour of the Centre for the Study of Existential Risk at Cambridge University puts it: “Governments need to think about security as just one category of risk.”
I'll write my comments in the comments so they can get up/downvoted appropriately.