What We Owe the Future

William MacAskill presents the case for longtermism: the view that positively influencing the long-term future is a key moral priority of our time. He opens with a thought experiment asking readers to imagine living, consecutively, through every human life ever lived, a total experience spanning nearly four trillion years. If that life were extended into the future, even a conservative estimate of humanity's remaining life span would mean that 99.5 percent of all experience still lies ahead. MacAskill contends that if we take seriously the idea that future people matter morally, that there could be vastly many of them, and that present actions can shape their world, then we face a responsibility of enormous scale. He introduces three recurring metaphors: humanity as a reckless teenager whose decisions carry lifelong consequences, history as molten glass that can be blown into many shapes but may soon cool and set, and the pursuit of long-term impact as an expedition into uncharted terrain.

MacAskill argues that concern for future people is common sense, citing the Iroquois Confederacy's "seventh-generation" principle, which instructs leaders to consider the welfare of descendants seven generations hence. He acknowledges that people may reasonably give extra weight to present generations but contends that these considerations do not override moral concern for the future. The potential scale reinforces this point: if humanity survives one million years, future people would outnumber those alive today ten thousand to one. MacAskill argues that the future could be extraordinarily good or extraordinarily bad, citing the resurgence of slavery during the colonial era and the rise of totalitarianism as reminders that moral regress is possible. He demonstrates that long-term impact is achievable, pointing to figures such as the Roman poet Horace, Shakespeare, and Benjamin Franklin, whose two-hundred-year charitable trust grew substantially over time. He presents decarbonisation as a proof of concept for longtermism: It reduces the 3.6 million annual deaths from fossil fuel air pollution while improving long-term climate outcomes.

MacAskill introduces a framework for evaluating long-term actions based on significance (how much value a change adds), persistence (how long it lasts), and contingency (whether the change would have happened otherwise). He supplements this with expected value theory, arguing that even modest probabilities of catastrophe justify significant action when the stakes are enormous. He also identifies "moments of plasticity," periods when institutions and technologies are still malleable, arguing that many current challenges represent such moments.

The book's central section addresses two ways to improve the long-term future: changing civilisation's trajectory and ensuring its survival. On trajectory, MacAskill uses the abolition of slavery as his primary case study, tracing how a small group of Quaker activists, led by the moral radical Benjamin Lay, helped catalyze one of history's most surprising moral transformations. Subsequent activists, including Anthony Benezet, Thomas Clarkson, William Wilberforce, and formerly enslaved people like Olaudah Equiano, built on this foundation. MacAskill argues that abolition was not economically inevitable: slavery was enormously profitable, and the British government paid £20 million to compensate slave owners. Drawing on historian Christopher Leslie Brown, who describes British abolitionism as "a historical accident" (68), MacAskill contends that multiple stable moral equilibria, meaning different self-reinforcing sets of norms a society could settle into, are plausible, and that value systems entrench themselves by suppressing competition.

MacAskill then argues that artificial general intelligence (AGI), a system capable of performing as wide an array of tasks as humans, could enable the permanent lock-in of values. He illustrates this with the Hundred Schools of Thought, a period of intense philosophical competition in ancient China. The Qin dynasty's attempt to lock in Legalism, a doctrine of strict law and centralized state control, lasted only fifteen years, while the Han dynasty's adoption of Confucianism, a tradition emphasizing ethical conduct and social harmony, shaped Chinese life for over a thousand years. AGI could extend such lock-in indefinitely because software is replicable and potentially immortal. MacAskill warns that computational analysis and expert surveys suggest a greater than 10 percent chance of AGI by 2036, and that attempts to align AGI with human values carry a grave risk of losing control to the systems themselves. Rather than locking in current values, he proposes a "morally exploratory world" structured so that better norms prevail over time through cultural diversity, political experimentalism, and free speech.

On ensuring civilisation's survival, MacAskill identifies three threats. The first is direct extinction, where he focuses on engineered pathogens made possible by increasingly powerful biotechnology. He documents a disturbing record of laboratory accidents, including three UK foot-and-mouth outbreaks (two in 2007 from the same lab), a 1979 Soviet anthrax leak that killed over a hundred people, and the fact that the last person to die of smallpox, Janet Parker in 1978, was infected by a lab leak. He also highlights great-power war as a catalyst for other catastrophic risks. The second threat is civilisational collapse, which global civilisation could likely survive even in extreme scenarios, though fossil fuel depletion could jeopardize recovery: Lewis Dartnell, a researcher who has studied civilisational recovery, concluded that an industrial revolution without coal would be "at a minimum, very difficult" (139). The third is technological stagnation. Progress inherently gets harder, and the growth in researcher numbers that historically compensated is ending as global population peaks. A civilisation capable of engineering pathogens but unable to defend against them faces constant danger.

MacAskill addresses the philosophical question of whether the nonexistence of future generations constitutes a moral loss. Drawing on the work of Derek Parfit, who inaugurated the field of population ethics, he argues that it does, provided those people would have had sufficiently good lives. He favors the "total view," in which more total wellbeing makes a population better, while acknowledging its uncomfortable implication: the "Repugnant Conclusion," which holds that a sufficiently enormous population with barely positive wellbeing could be better than a smaller population of flourishing lives. He then examines whether the future is likely to be good, estimating that roughly 10 percent of the global population currently has below-neutral wellbeing, though the picture for nonhuman animals is far grimmer, with billions killed for food annually after lives of intense suffering. MacAskill argues for cautious optimism based on an asymmetry: people frequently pursue good things because they are good but rarely pursue bad things simply because they are bad.

MacAskill concludes with practical guidance. He argues that donations to effective nonprofits are far more impactful than personal consumption changes: a $3,000 donation to the Clean Air Task Force, a nonprofit focused on clean energy and emissions reduction, would reduce carbon emissions by an expected three thousand tonnes per year, dwarfing the roughly sixty-four-tonne lifetime effect of going vegetarian. He urges readers to choose careers strategically and advocates for a community-based approach through the effective altruism movement, a community that uses evidence and reason to do the most good. He stresses that longtermism should supplement commonsense morality rather than replace it. In a preface added to the paperback edition, MacAskill refines his definition of longtermism to "we should do much more to protect the interests of future generations" and highlights urgent developments including Russia's nuclear threats and the rapid advance of AI. The epilogue identifies specific priority areas including biosecurity, nuclear command-and-control improvements, clean-energy innovation, AI interpretability research (efforts to understand what AI systems are doing internally), protecting democracy, and supporting talented youth in lower-income countries.