Episode Summary
Executive Summary: Joshua Greene explains how moral psychology reveals the tug-of-war between emotional intuition and reason in dilemmas like the trolley problem, then applies that insight to charitable giving, AI decision-making, and political polarization. He argues that small design changes can steer people toward more effective donations and more cooperative cross-group interactions.
Main Topics: Trolley problem as a model of moral judgment (Priority: 5/5): Greene uses the switch vs. footbridge versions to show how direct, intentional harm triggers stronger emotional responses than more indirect harm, illuminating competing moral processes in the brain. Emotion, reason, and dual-process moral psychology (Priority: 5/5): The discussion frames moral choice as a dual-process system: automatic emotion/intuition versus deliberative reasoning, with examples from psychopaths and Buddhist monks. Effective altruism and charitable giving design (Priority: 5/5): Greene explains huge variation in charity effectiveness and how Giving Multiplier encourages split donations and matching funds to reconcile personal preference with evidence-based impact. GiveDirectly and cash transfers (Priority: 4/5): The episode highlights direct cash transfers as a highly effective and autonomy-respecting way to help people in extreme poverty, including a podcast campaign for Rwanda. AI, self-driving cars, and moral trade-offs (Priority: 4/5): Greene argues trolley-style thinking is limited but still useful for training AI around risk, constraints, and acceptable trade-offs in driving and other automated decisions. Reducing polarization through cooperative contact (Priority: 5/5): Greene describes Tanko, a cooperative quiz game designed to build trust across political divides through mutually beneficial teamwork rather than debate or persuasion.
Key Arguments: Moral judgments are shaped by both emotion and reason; the footbridge case reliably provokes stronger emotional aversion because it involves direct, intentional, personal harm. Psychopaths and some Buddhist monks may endorse pushing the man for different reasons: reduced emotional aversion in psychopaths, and trained detachment plus greater-good reasoning in monks. Charitable effectiveness varies enormously; some interventions are 100x to 1,000x more impactful per dollar than others, so donation choices should consider impact, not only emotional closeness. People usually prefer emotionally meaningful local charities, but a split-donation model lets them satisfy personal attachment while also directing money to highly effective causes. Giving Multiplier increases effective giving by adding matching funds when donors split between a favorite charity and a top-rated charity, making the choice psychologically appealing. Cash transfers via GiveDirectly are powerful because they preserve recipient agency and can stimulate local economies while meeting immediate needs. For AI and self-driving cars, the goal should primarily be reducing fatalities and injuries; trolley-style dilemmas matter mostly as training frameworks for risk thresholds and constraints. Polarization is best reduced not by abstract dialogue alone but by structured, enjoyable cooperation that makes people depend on one another to succeed. Tanko works because it creates a positive, mutually beneficial task that is scalable, enjoyable, and not limited to already-open-minded participants. Small interventions in psychology can produce real-world impact when they are designed around how people actually make choices.
Data Points: Footbridge approval among Buddhist monks: 80% - In an unpublished senior-thesis study in Tibet, 80% of 50 Buddhist monks approved pushing the man off the footbridge. Participants in monk study: 50 monks - Sample size in the Tibet study described by Greene. Effectiveness difference in charities: 100x to 1,000x - Greene says some charities can be dramatically more effective than others when measured by lives saved or quality of life improved per dollar. Seeing-eye dog training cost: about $50,000 - Example of a U.S. blindness-related intervention versus more cost-effective trachoma surgery abroad. Trachoma surgery cost: less than $100 - Simple surgery to prevent blindness in places where trachoma is prevalent. Malaria net cost: about $5 per net - Cost to distribute an insecticidal bed net in sub-Saharan Africa. Lives saved by malaria nets: about 1 life per 1,000 nets - Greene cites this as a rough estimate for net distribution in key areas. Cost per life saved via malaria nets: around $5,000 - Derived from the net distribution estimate. De-worming treatment cost: less than $1 - Example of a highly cost-effective global health intervention. Giving Multiplier total raised: over $5 million - Amount raised since the platform launched five years ago. Giving Multiplier amount to super-effective charities: close to $3 million - Portion of funds directed to highly effective charities through the platform. Giving Multiplier donors: over 5,000 donors - Total number of donors using the platform. Giving Multiplier favorite-charity users: over 2,600 people - Donors who gave to their personal favorite charities through the platform. Estimated lives saved via GiveWell-aligned health and poverty charities: about 220 lives - Greene’s estimate of impact from money moved through Giving Multiplier. Control condition preference for favorite charity: about 80% - In experiments, most people chose their favorite charity over the super-effective option when given a direct choice. Split-option choice rate: a little over 50% - When offered a 50-50 split option, more than half chose it in the first experiment. Increase from matching funds: another 70% of people - Adding money on top of the split increased uptake substantially. Positive campaign match: 50% on top - Giving Multiplier adds up to 50% extra when donors give to its super-effective list. Podcast campaign goal: $1 million - Pods Fight Poverty campaign goal for GiveDirectly. Villages targeted in Rwanda: 3 villages - Campaign aims to support three villages in Rwanda. Families targeted in Rwanda: over 700 families - Planned beneficiaries of the GiveDirectly campaign. Economic boost from cash transfers: 2.5x - Recent studies cited by Greene suggest each dollar transferred can generate a 2.5x boost in the local economy. U.S. road fatalities: 40,000 - Greene cites annual road deaths in the context of self-driving car safety. Tango immediate effect: about 15 years of animosity rolled back - Feeling-thermometer change after playing the cooperative game. Tango retention period: up to 4 months - Effects persisted in five randomized controlled trials. Tango enjoyment rating: 10/10 (median/modally reported) - Participants rated the game extremely highly. Harvard incoming class size: over 1,000 students - Used in a campus rollout of Tango for the class of 2029. Harvard political composition: about 80% liberal, 20% conservative - Context for the Harvard pilot study. Harvard liberal warmth increase: 7 points - Liberals became warmer toward conservatives after playing Tango. Harvard conservative comfort increase: 5 points - Conservatives felt more comfortable voicing controversial opinions at Harvard after the game. Participants who met in person after Tango: over 10% - Teams who chose to connect after the game met up in the dining hall.
Pivotal Quotes: "The gist of it is that these moral dilemmas are nice little fruit flies for understanding distinct, dissociable, and sometimes competing mechanisms in our brain." — Joshua Greene: Explaining why the trolley problem matters for studying moral psychology. "The people who are the most drawn to is typically not the stuff that's most effective." — Joshua Greene: Describing the core tension in charitable giving between emotional preference and impact. "We want to get people on the same team and do it in a way that is scalable so you can really make a dent in the problem." — Joshua Greene: Explaining the design goal behind Tango and cross-partisan cooperation research.
Implications: The episode suggests better outcomes come from designing choices around human psychology: make effective giving emotionally satisfying, build AI with risk-aware moral constraints, and scale cooperation through fun, shared tasks rather than persuasion alone.