Skip to main content

Between about 2015 and 2022, many of the movement's leading thinkers shifted their attention from present poverty toward the distant future. The shift began with a thought experiment from the British philosopher Derek Parfit.

Parfit's three outcomes

At the end of Reasons and Persons (1984), Parfit asked readers to compare three outcomes: peace; a nuclear war that kills 99 percent of the world's people; and a nuclear war that kills everyone. Most people think the greater difference lies between peace and the first war. Parfit disagreed. Parfit argued that killing everyone would be far worse than killing 99 percent, because it would destroy all future generations. The survivors of the first war could rebuild, and their descendants might live for millions of years. The second war would close off that whole future.

Three premises

In What We Owe the Future (2022), William MacAskill defended longtermism, which he described, in paraphrase, as the view that making the long-term future go well is a key moral priority of our time. The case rests on three claims.

The first premise is that future people matter morally just as much as people alive now. Where Singer said distance in space does not lessen a person's claim, longtermists say distance in time does not either.

The second premise is that there could be enormous numbers of future people. A typical mammal species survives around a million years, and our own has existed for only a few hundred thousand. If humanity lasted even that long at today's population, future people would outnumber the living by thousands to one, and if people one day settled beyond Earth the numbers would be far greater.

The third premise is that we can make the future go better. This is the hardest claim to establish, and most of the debate turns on it.

Existential risk

The clearest way to affect the far future is to make sure it happens at all. An existential risk is one that threatens the destruction of humanity's long-term potential, whether through extinction, an unrecoverable collapse of civilization, or a permanent and terrible tyranny.

In The Precipice (2020), Toby Ord gave rough estimates of these risks. He put the total chance of an existential catastrophe within the next hundred years at about one in six. Ord judged risks created by human action to be far greater than natural ones. He put the chance of a catastrophe from all natural causes, such as asteroids and supervolcanoes, at about one in ten thousand per century, but the chance from engineered pandemics at about one in thirty and from artificial intelligence whose goals diverge from ours at about one in ten. Ord stressed that such figures are judgments, not measurements, offered to make his reasoning open to challenge.

Astronomical stakes

The philosopher Nick Bostrom argued in 2003 that, given the vast numbers of possible future lives, even a small delay or reduction in humanity's prospects would be an astronomical loss. Hilary Greaves and MacAskill later defended a stronger thesis. Strong longtermism holds that in many of the most important choices we face, the effects on the far future matter most. The reasoning is expected value at enormous scale: if trillions of future lives hang on our choices, even a tiny reduction in the risk of extinction could outweigh large benefits today.

From theory to funding

Money followed the argument. Funders in the movement gave large sums to biosecurity, research on the safety of artificial intelligence, and the reduction of nuclear risk. Longtermists point out that this work is not as remote as it sounds. Many measures against existential risk, such as preparing for pandemics, also protect people alive now. Ord's book was published in March 2020, as COVID-19 spread around the world, and its warnings about pandemic preparedness suddenly seemed less abstract.

Worked example: watching the sky

Asteroid impacts are the model case for existential risk, because they are well understood. An object about ten kilometers wide, like the one linked to the extinction of the dinosaurs 66 million years ago, strikes Earth very rarely, on the order of once in a hundred million years. In 1998 the United States Congress directed NASA to find most near-Earth asteroids larger than one kilometer, and within about a decade and a half the survey had found more than 90 percent of them. None is on course to hit us this century. In 2022 NASA's DART spacecraft struck a small asteroid moon called Dimorphos and measurably changed its orbit, showing that deflection can work.

Now run the longtermist calculation. Suppose that before the survey there was a one in a million chance of a civilization-ending impact this century, and the survey and follow-up work cut that chance in half. If the future holds trillions of lives, that small change in probability saves, in expected value, millions of lives, at a cost that was modest by the standards of space budgets.

The asteroid case is easy because the risk can be measured and the fix is clear. The harder cases, such as artificial intelligence or engineered disease, involve probabilities that are little more than informed guesses. Longtermism's strength is to take such risks seriously; its weakness is that the arithmetic can run far ahead of the evidence, as its critics were quick to argue.

Quiz

5 questions · Practice: scored, not saved

1. What did Parfit argue about his three outcomes?
2. How did MacAskill describe longtermism, in paraphrase?
3. What is the first premise of the case for longtermism?
4. What is an existential risk?
5. How did Ord compare natural risks with risks from human action?
6. What total chance of existential catastrophe this century did Ord estimate?
7. What does strong longtermism hold?
8. What point do longtermists make about many existential-risk measures?

CivilSimian.com created by AxiomaticPanic, CivilSimian, Kalokagathia