#56 — Abusing Dolores

Episode Summary

Executive Summary: Sam Harris and Paul Bloom revisit Bloom’s critique of empathy, arguing that emotional contagion is biased, parochial, and often morally misleading, while compassion paired with reason is a better basis for ethics. They extend the discussion to family partiality, effective altruism, public policy, and the coming challenge of AI and robots that may appear human-like yet require morally coherent design.

Main Topics: Empathy vs. Compassion (Priority: 5/5): Bloom distinguishes empathic emotional contagion—feeling another’s pain—from compassion, which cares for others without sharing their suffering. He argues empathy is biased and can produce burnout, while compassion is pleasurable and more sustainable. Moral Bias and Scope Insensitivity (Priority: 5/5): The hosts discuss how empathy narrows concern to salient individuals and leads to irrational moral judgments, especially in charity and public policy. They cite research showing people respond more to single identifiable victims than to statistics or larger groups. Partiality, Family, and Moral Tradeoffs (Priority: 4/5): They debate whether it is morally acceptable to love one’s children, friends, and in-group more than strangers. Both acknowledge special obligations to family, while also emphasizing that fair institutions should not be built around personal bias. Effective Altruism and Automating Morality (Priority: 4/5): Sam Harris describes using automatic charitable donations and endorses systems like constitutions, blind review, and charity evaluators to bypass unreliable gut instincts and make better moral decisions at scale. Identity Politics, Conflict, and In-Group Empathy (Priority: 4/5): Bloom argues that many conflicts are fueled by empathy concentrated within groups, not by a lack of empathy overall. He suggests identity politics and political polarization are intensified by selective emotional allegiance. AI, Robots, and Moral Programming (Priority: 5/5): The conversation shifts to how empathy and compassion should be encoded in AI. Bloom argues machines should be designed with rational compassion and consistent moral rules, not emotional contagion, which could make them capricious or racist. Conscious Machines and the Ethics of Treatment (Priority: 5/5): They explore whether humanoid robots will be treated as conscious by default, and whether abusing human-like machines would damage human morality even if the machines are not sentient. Westworld and Ex Machina are used as philosophical thought experiments.

Key Arguments: Empathy, understood as feeling another’s pain, is morally unreliable because it is selective, biased toward in-groups, and often worsens suffering for the empathizer. Compassion is different from empathy: it is caring for others and wishing them well without becoming emotionally overwhelmed by their suffering. Moral judgment should rely more on reasoning, cost-benefit analysis, and stable principles than on vivid stories or emotionally charged images. People systematically misallocate concern: identifiable victims trigger more help than statistical victims, even when many more lives are at stake. Partiality toward one’s children and loved ones is normal and arguably appropriate in personal life, but institutions should be designed to counteract those biases. Effective altruism offers practical tools—charitable evaluation, automation, pledges—to convert moral insight into consistent action. AI should not be programmed to mirror human emotional bias; if it is to make consequential decisions, it should embody rational compassion and coherent principles. Human-like robots may be treated as conscious regardless of philosophical uncertainty, which raises concerns about digital slavery and the moral degradation of abusing them. If consciousness can arise from information processing, then creating suffering machines would be a major moral catastrophe. Policy and technology choices always encode values; there is no neutral default, only hidden moral decisions. In-group empathy can intensify conflict because each side feels maximal concern for its own and little for the other side. The best safeguards against our own irrationality are institutional and procedural: constitutions, rules, blind review, and automated giving.

Data Points: Podcast appearance count: 3rd appearance - Sam Harris notes Paul Bloom is his only guest to appear three times, making him a standout return guest. Effective altruism donation benchmark: $3,500 - Harris cites a claim that roughly $3,500 donated to highly effective charities can save one human life. Malaria burden reduction: 50% - Harris references the Against Malaria Foundation and says malaria deaths have come down by about half. Annual malaria deaths: close to 1 million - Used to illustrate the ongoing scale of preventable global suffering. Annual malaria deaths in recent past: 2 million - Harris notes the death toll was very recently around 2 million per year. Speaker's charitable automation: first $3,500 each month - Harris says the first $3,500 of podcast revenue each month goes automatically to the Against Malaria Foundation. Autonomous car safety statistic: tens of thousands - Harris notes that tens of thousands of people die each year in car crashes, motivating self-driving car ethics. Autonomous car safety improvement: factor of 10 or 100 - He suggests self-driving technology could reduce deaths by one to two orders of magnitude. Meditation retreat duration: 6 months each year - Harris says Matthew Ricard spent half the year in total solitude when Harris met him. Solitary practice duration: more than 20 years - Harris describes a lama who practiced in solitude for over two decades.

Pivotal Quotes: "empathy is a wonderful source of pleasure... But what I argue is in the moral realm, when it comes to being good people, it steers us dangerously astray" — Paul Bloom: Bloom summarizes his thesis that empathy is useful in some domains but harmful as a moral guide. "what we should replace empathy with for decision making is cold-blooded reasoning of a more or less utilitarian sort where you judge costs and benefits" — Paul Bloom: Bloom explains his preferred framework for moral judgment and policy. "cold and bureaucratic is much better than hot-blooded and arbitrary" — Sam Harris: Harris argues that institutions and policy should rely on procedures rather than emotional improvisation.

Implications: Listeners are urged to favor rational compassion over emotional identification, and to design institutions and AI systems that reduce bias, not amplify it. The future of ethics and technology will depend on building rules that outperform intuition.