In one sentence: What We Owe the Future is Oxford philosopher William MacAskill’s case for longtermism, the idea that because future people count, could vastly outnumber us, and can be helped or harmed by what we do now, positively shaping the long-run future, by guarding against extinction, civilizational collapse, stagnation, and the permanent lock-in of bad values, is one of the great moral priorities of our time.
At a Glance
Author: William MacAskill
First published: 2022 (Basic Books)
Category: Philosophy / Society & Current Affairs
Length: 352 pages, about 120,000 words including extensive endnotes and appendices
ISBN-13: 9781541618626 (Basic Books hardcover)
Summary reading time: about 13 minutes
Book reading time: about 8 hours, roughly 6 for the main text
Notable adaptations: none
William MacAskill is an associate professor of philosophy at Oxford and a founder of the effective altruism movement, which asks how to do the most good with the evidence available. He cofounded Giving What We Can, whose members pledge at least 10 percent of income to effective charities, and 80,000 Hours, which advises on high-impact careers. He gives away everything he earns above a modest threshold. This book, published with a large publicity push in 2022, is the movement’s most ambitious statement, extending its logic from present-day poverty to the whole of the human future.
Read it if you enjoy big-canvas moral argument grounded in history and probability, and want to understand the longtermist worldview that now shapes real funding and policy debates around AI, pandemics, and existential risk. Skip it if you have no patience for expected-value reasoning about distant, uncertain outcomes, because that machinery drives the whole book.
The Big Idea
MacAskill’s argument fits in three sentences of his own: “Future people count. There could be a lot of them. We can make their lives go better.” Morally, he argues, distance in time is like distance in space. Just as a child cut by glass you dropped matters whether she is cut next week or next century, harm is harm whenever it occurs. And the numbers are staggering: if humanity lasts merely as long as a typical mammal species, the people yet to be born outnumber us thousands to one, which means that on any view where people matter equally, the future is where most of the moral stakes live. Strange as it sounds, we may be the ancients, living at the very beginning of history.
The practical question is whether we can actually help people that distant. MacAskill’s answer is that we live at an unusual hinge: a moment of rapid technological change, unprecedented global connection, and unusually plastic values, after which things may harden. He offers two levers. We can make the future longer, by reducing risks of extinction, unrecoverable collapse, and stagnation. And we can make it better, by improving the values that get locked in, because ideas, like glass, are molten only briefly and then set for a very long time.
Key Ideas
1. Future people count, and there could be an astonishing number of them
The book opens with a thought experiment: imagine living, in order, every human life ever lived, roughly 100 billion so far, and then every life to come. On mundane assumptions about how long species last, more than 99 percent of that composite life still lies ahead. MacAskill’s point is not science fiction but arithmetic applied to ethics. Future people are real people who will love, suffer, and hope, and they are completely disenfranchised: they cannot vote, lobby, or tweet in their own defense. He frames longtermism as the next step in the expansion of the moral circle, extending consideration to those without a voice, and he is careful about the claim’s strength. The book argues that the long-term future is a key priority, not the only one, and MacAskill continues to fund present-day global health work. But once you accept that time of birth is morally arbitrary, the sheer scale of what could come makes the future very hard to ignore.
2. History can be steered, if you catch it while the glass is molten
Why think anyone can affect a millennium? MacAskill’s tool is the SPC framework: an action’s long-run value is the product of the significance of the change it makes, the persistence of that change, and its contingency, meaning how unlikely it was to happen anyway. His central historical case is the abolition of slavery. Against the view that abolition was economically inevitable, he follows historians who argue it was, in one scholar’s words, a historical accident: slavery remained profitable, Britain paid enormous compensation and enforcement costs to end it, and morally arbitrary practices persist elsewhere for centuries. If a handful of Quaker moral entrepreneurs like Benjamin Lay genuinely changed the world’s values, then values are contingent, and campaigners today may matter just as much. The other key concept is plasticity: like molten glass, institutions and ideologies are malleable early and rigid later. Korea’s border, drawn overnight by two colonels, has now lasted eighty years. The window when you can shape a new technology, constitution, or norm is brief, and it is usually open right before the change looks important.
3. The biggest trajectory risk is locking in bad values, and AI could do it permanently
The book’s most original chapter argues that the worst outcomes short of extinction involve value lock-in: a single ideology or regime entrenching itself beyond challenge. History offers rehearsals, from Qin China’s book burnings to the Han dynasty’s thousand-year installation of Confucian orthodoxy. What makes the coming century different is artificial general intelligence. Software, unlike emperors, is potentially immortal: it can be copied perfectly, survives its hardware, and could concentrate unprecedented economic and military power. Writing gave ideas the power to influence society for thousands of years, MacAskill observes, and artificial intelligence could give them influence that lasts millions. He worries both about AI takeover, systems pursuing goals misaligned with ours, and about humans using aligned AI to entrench their own values forever. Since he thinks it is extraordinarily unlikely that our current values are the best possible, the goal is not to lock in today’s best guess but to keep moral evolution going: preserve pluralism and experimentation, protect good-faith debate, and delay irreversible commitments until humanity has had time to reflect. He concedes the paradox that even an open, exploratory world requires locking in some norms, like tolerance itself.
4. Extinction, collapse, and stagnation are the three ways to lose the future outright
The survival half of the book works through three failure modes. Extinction: MacAskill treats engineered pandemics as the clearest near-term danger, worse than asteroids by orders of magnitude, given the history of state bioweapons programs and repeated lab accidents, with nuclear war and misaligned AI close behind. Collapse: here he is surprisingly reassuring, since civilization has rebuilt from catastrophes before, but he flags one ratchet, fossil fuel depletion. Easily accessible coal and oil powered the first industrial revolution, and if a collapsed civilization had already burned them, reindustrialization could be far harder, which makes decarbonization what he calls a proof of concept for longtermism, a win for the present and an insurance policy for any future restart. Stagnation: drawing on evidence that ideas are getting harder to find and that global fertility is falling, he argues technological progress could stall for centuries, which would be dangerous rather than merely disappointing, because it would strand humanity at a level of destructive capability, engineered pathogens, nuclear arsenals, without the tools to defend against them. A climber halfway up a cliff face cannot safely stop.
5. Whether the future is worth saving is a real question, and his answer is a sober yes
Unusually for a book in this genre, MacAskill spends two chapters asking whether more future is actually good. That requires population ethics, philosophy’s most treacherous territory. He walks through the standard views and their bullet-biting costs, including Derek Parfit’s Repugnant Conclusion, and lands, without full confidence, near the total view: more good lives make the world better. Then he asks about the actual sign of the future. Surveys suggest most people today report lives worth living, and the world is getting richer and happier. But he refuses to look away from the ledger’s other side: humanity farms and kills tens of billions of land animals a year in conditions he judges decisively negative, enough that combined human and animal wellbeing today may be net bad. His case for optimism rests on an asymmetry of motivation: catastrophically bad futures generally require agents optimizing for badness, which is rare, while good futures have many paths, because people often do good simply because it is good. Hope, in his telling, is a probability judgment, not a temperament. “There’s no inevitable arc of progress,” and no rescue coming: “It’s on us. And we are not destined to succeed.”
6. Uncertainty is not an excuse, it is the operating condition
The book’s decision-making advice is built for fog. Borrowing the Aymara image of walking backward into the future, facing the known past, MacAskill offers three rules for acting under deep uncertainty: take actions that are robustly good across many scenarios, preserve and build options, and invest in learning more. He pairs the SPC framework with importance, tractability, and neglectedness for choosing causes, which is why he emphasizes areas like AI safety and pandemic preparedness, at the time tiny fields, over already-crowded ones. His probabilities are explicit and explicitly provisional, and he volunteers, disarmingly, that he expects to disagree with some of his own numbers soon. On the deepest objection, that this is all speculation multiplied by big numbers, his reply is that the probabilities he relies on, like a greater than one in a thousand chance of catastrophe this century, are not Pascalian tiny, and that uncertainty cuts both ways. We routinely buy insurance against far smaller risks to far smaller stakes.
7. What one person can actually do, and why movements beat lifestyles
The closing chapter is practical and deliberately deflationary about consumer virtue. Personal consumption choices matter far less than people assume, and MacAskill calls the fixation on them a strategic blunder: going vegetarian for life averts around 64 tonnes of carbon, while a well-aimed annual donation can plausibly avert thousands of tonnes every year. The big levers are donations to effective organizations, career choice, treating your working life as an experiment and your comparative advantage as the question, political participation, where even one vote’s expected value is surprisingly large, having and raising children, which he defends as a positive contribution to the future, and movement building, since persuading others multiplies everything else. He is emphatic that longtermism supplements rather than replaces common-sense morality: no ends-justify-means shortcuts, keep promises, tell the truth, stay kind. The book closes by pointing at the reader: history’s biggest changes came from small groups who took a moral idea seriously early, and “if there’s ever anyone who will take action on behalf of future generations, it’s you.”
Context and Analysis
What We Owe the Future arrived in 2022 as the flagship statement of longtermism and immediately attracted both acclaim and hard criticism, and both deserve airing. Philosophically, the book’s foundations in population ethics remain genuinely unsettled, which MacAskill admits: he leans on a total view whose Repugnant Conclusion he accepts, and critics argue the whole edifice rests on expected-value reasoning about probabilities that are, in places, constructed rather than measured, with pivotal numbers coming from a small circle of allied researchers rather than peer-reviewed consensus. Critics from the left call longtermism a license to deprioritize the suffering poor of the present, though the book explicitly denies the strong version of that claim, and MacAskill’s own career includes founding organizations focused on present-day global health. The sharpest real-world blow came months after publication, when FTX founder Sam Bankman-Fried, the movement’s largest donor and a figure MacAskill had advised, collapsed into fraud, an episode that damaged effective altruism’s reputation and gave the book’s warnings about ends-justify-means reasoning an unintended irony, though the fraud postdates the book and its arguments stand or fall on their own. What has aged best is the risk agenda: within two years of publication, AI safety went from the book’s neglected cause to a mainstream governmental concern. Read it as a serious, self-questioning argument whose conclusions you should discount by your own credence in its probabilities, which is exactly how its author says to read it.
On this site, the obvious companions are our summaries of Life 3.0, Max Tegmark’s map of possible AI futures that fills in the scenarios MacAskill compresses, and Human Compatible, Stuart Russell’s account of the AI alignment problem that underpins this book’s most alarming chapter.
How to Apply It
- Apply the SPC test to your biggest commitments. For a project, cause, or career move, ask how significant the change is, how long it will persist, and whether it would have happened anyway. Persistence and contingency are where most plans quietly score zero.
- Move money before lifestyle. If you care about a problem, a regular donation to an effective organization working on it will usually dwarf the impact of your consumption choices. Consider a fixed percentage of income, the Giving What We Can pledge starts at 10 percent.
- Treat your career as your largest grant. You will spend around 80,000 hours working. Ask where your comparative advantage lies against what the world neglects, and be willing to spend real time searching, since finding a path twice as impactful is worth a long search.
- Favor robustly good actions under uncertainty: build options, learn more, and avoid irreversible moves, in your own decisions as much as in policy.
- Watch for molten glass. In your field, notice which norms, standards, or technologies are still plastic, because influence is cheapest right before things set.
- Vote and engage politically in proportion to the real expected value, which is higher than cynicism suggests, especially where outcomes are close.
- Keep common-sense morality as a constraint. The book’s own warning, written before its movement needed it, is that calculating do-gooders who cut ethical corners destroy the very trust their causes run on.
Memorable Lines
“Future people count. There could be a lot of them. We can make their lives go better.” (William MacAskill)
“If humanity survives to even a fraction of its potential life span, then, strange as it may seem, we are the ancients: we live at the very beginning of history, in the most distant past.” (William MacAskill)
“Harm is harm, whenever it occurs.” (William MacAskill)
“Writing gave ideas the power to influence society for thousands of years; artificial intelligence could give them influence that lasts millions.” (William MacAskill)
“There’s no inevitable arc of progress.” (William MacAskill)
“It’s on us. And we are not destined to succeed.” (William MacAskill)
Should You Read the Full Book?
Verdict: Recommended
The full book earns its length for anyone who wants to engage seriously with longtermism rather than react to its headlines: the historical case studies, abolition, Qin and Han China, the fossil fuel ratchet, are the argument’s real substance, and the population-ethics chapters are among the clearest introductions to that maze ever written for a general audience. Read it whole if you work anywhere near AI, biosecurity, philanthropy, or policy, or if you want your objections to be aimed at what MacAskill actually says, which is more hedged and more self-critical than either fans or detractors suggest. If you mainly want the worldview, the three-sentence case, the SPC framework, the lock-in warning, and the donations-over-lifestyle conclusion, this summary carries it, and the book’s probability appendices will not change your mind if the method already leaves you cold.
The What We Owe the Future book page has the full details and where to get a copy.