Singer, Peter

professor of bioethics; author of "Animal Liberation" and "The Life You Can Save"; key popularizer of the effective altruism movement

Peter Singer: Suffering in Humans, Animals, and AI

Lex Fridman Podcast · Jul 8, 2020

• Peter Singer
• utilitarianism
• animal rights
• speciesism
• effective altruism
• suffering
• AI consciousness
• existential risk
• moral philosophy
• The Life You Can Save

Core ideas

Singer traces his early awareness of suffering to losing three of his four grandparents in the Holocaust, and reflects that WWII's legacy included both a beneficial delegitimizing of overt racism in the postwar West and a background dread of nuclear annihilation that shaped his childhood. He rejects the idea that war or shared suffering has redeeming value — while acknowledging that adversity can forge bonds and even wartime "brotherhood," he argues equivalent closeness is achievable without suffering or death, and is skeptical of the unquestioning patriotism that group suffering often produces. He defines suffering as a conscious state that, considered in itself, one would rather not have or repeat — explicitly not denying that suffering can be chosen for good reasons (a marathon, a dentist visit) or have positive effects (creativity, resilience), but insisting that, all else equal, it is something to minimize. He distinguishes objective suffering (hunger, cold, chronic pain — the target of his practical, present-day ethical focus) from relative suffering (boredom, lack of purpose, which can emerge once basic needs are met in affluent societies) and says he would only turn attention to the latter once the former is largely eliminated; he does not believe suffering will ever be fully eradicated, and floats the idea that eliminating it entirely might also eliminate life's highest positive experiences.

On human nature, Singer resists claiming everyone is capable of evil, but argues most people are highly susceptible to environmental and situational pressure, guessing fewer than 10% of people could truthfully claim they would have resisted participating in Holocaust atrocities under Nazi indoctrination and coercion — and says he cannot confidently place himself among that 10%. He connects this to his own experience becoming a vegetarian more than 40 years ago, at real but modest social cost, as an example of taking an ethical stand when the "right" action is inconvenient. Singer's central animal-ethics idea, developed in "Animal Liberation," is "speciesism" (a term he borrowed from Richard Ryder but sharpened philosophically): a prejudice against the interests of non-human beings analogous to racism or sexism, whereby the significance of pain and suffering is wrongly treated as dependent on species membership — a bias, traceable through Aristotle and Kant, that he holds responsible for factory farming, non-essential animal research, and animal use in sports and entertainment. He explicitly disagrees with Kant's and Aquinas's rationale for treating animals well (that cruelty to animals might make us more likely to be cruel to humans), insisting instead that animal suffering matters directly and equally regardless of species.

On consciousness and rights, Singer argues robots should have rights once (and only once) they are capable of consciousness — having a genuine subjective perspective from which life can go well or badly — and that consciousness is a prerequisite for suffering, though not strictly identical to it (he speculates a system could in principle be built to feel pleasure while being incapable of suffering). He distinguishes current AI (like Google Maps) as clearly non-conscious from a hypothetical future system that might display humor, wit, fear, or distress in ways we cannot mechanistically verify or falsify — arguing that in cases of serious uncertainty (as with insects, compared to vertebrates where we can be more confident) we should extend the "benefit of the doubt," while pushing back against the idea (raised by the interviewer) that a convincingly faked display of suffering is itself morally equivalent to real suffering; he likens this view to mid-20th-century philosophical behaviorism (Gilbert Ryle, Norman Malcolm), which he considers largely discredited, while granting that responding protectively to apparent suffering displays may still be psychologically healthy insofar as it keeps people from hardening themselves against real suffering elsewhere. Asked what he'd want to ask a hypothetical AGI to probe for consciousness, he proposes asking it to describe its felt experience directly, and separately testing its sense of persistent self-identity (e.g., via a brain-transplant thought experiment) — and speculates that if digital minds can be paused and restored on new hardware, "suffering" and identity may ultimately be substrate-independent, analogous to how human memory might one day be technologically augmented (he expresses enthusiasm about brain-computer interfaces expanding memory access).

On AI existential risk, Singer says he doesn't know what to think, personally, when weighing warnings from thinkers like Nick Bostrom and Toby Ord against reassurances from practitioners at organizations like DeepMind that human-level AGI is not imminent — but affirms that "eventually" this could become a real problem worth researching now, even at low probability, given how catastrophic human extinction would be. He resists treating existential risk as carrying literally infinite disutility in ethical calculations (noting even Bostrom's own estimates of potential future digital lives, while vast, are finite), and flags as a genuinely disputed premise the idea that the loss of merely possible future lives is straightforwardly comparable to the suffering or deaths of people who actually exist. Explaining utilitarianism, Singer defines it as the view that the right action is the one that produces the best consequences (weighted by probability), with classical utilitarianism cashing out "best" as maximizing happiness and minimizing suffering; he acknowledges that in practice measuring and comparing quantities of happiness and suffering, and predicting long-term consequences, remains extremely difficult, so utilitarian reasoning functions mainly as a rough guide rather than a precise calculus. He rejects a symmetric view of pleasure and pain (arguing the worst possible suffering is farther below neutral than the best possible pleasure is above it, illustrated by his own unwillingness to trade even many hours of extreme pain for a single hour of his most pleasurable experience), which leads him to resist naive aggregation that would sacrifice a few people's extreme suffering for large aggregate gains to others — while still ultimately endorsing aggregation as correct at some threshold, even where it conflicts with intuitions about justice and fairness.

On solving disagreement about ethics and policy, Singer defends democracy despite its flaws as preferable to any realistic alternative (including a hypothetical benevolent dictator, given the empirical tendency of concentrated power to corrupt), while doubting that public consensus could ever converge on a single utilitarian "objective function" for policy given persistent, legitimate disagreement about competing values like justice and priority for the worst-off. He describes effective altruism's practical core (developed in "The Life You Can Save," free online) as directing giving toward independently verified, highly effective charities, recommends resources like 80,000 Hours for choosing high-impact careers (including "earning to give" via lucrative careers), and lays out his own progressive suggested giving scale (from 1% of income at modest levels up to a third at very high incomes) as a non-burdensome, ethically motivated minimum — while also arguing, citing psychological research, that such giving tends to increase rather than diminish life satisfaction and provides a durable sense of purpose. Reflecting on mortality in his 70s, Singer states he does not believe in an afterlife or mind-uploading rescuing him personally, finds his philosophy offers no special comfort on this point beyond acknowledging that finite time motivates action, and closes by describing life's meaning as something humans create rather than discover: since some things (extreme suffering) are objectively bad and others (a rich, fulfilling life) are objectively good, meaning consists in doing what one can to increase the good and reduce the bad.

Excerpts