We are taught, from childhood, that empathy is the bedrock of morality. Feel what others feel. Walk in their shoes. The more empathy, the better the person. It sounds unassailable. It is also, according to a growing body of evidence, dangerously wrong.

Psychologist Paul Bloom spent years arguing precisely this. His conclusion, backed by experimental research rather than contrarianism alone: empathy is a terrible moral guide. Not because caring is bad, but because empathy — the visceral, emotional sharing of another person’s experience — is irrational, biased, and routinely leads to decisions that harm far more people than they help.

The Spotlight Problem

Empathy works like a spotlight. It illuminates one person, one face, one story with extraordinary intensity. The problem is that moral decisions rarely involve one person.

In 1983, a terminally ill girl named Jessica McClure fell into a well in Texas. Millions of dollars poured in for her rescue. Simultaneously, thousands of children were dying of preventable diseases in conditions far more desperate — and received almost nothing. The same dynamic plays out constantly. Charities that feature a single named child raise dramatically more money than those presenting statistical data about thousands. Donate $100 to save one identified child, or $10 to a program that saves fifty anonymous ones. Most people choose the one. Empathy chose for them.

The spotlight doesn’t just narrow focus — it distorts priorities entirely. This has a direct parallel in how we perform social virtues generally: see kindness has become emotional labor for how the performance of care can actively crowd out care itself.

Empathy Is Biased by Design

Empathy is also profoundly, structurally unfair. We feel it most intensely for people who resemble us — in appearance, background, proximity. Studies consistently show that we empathize more readily with those who share our ethnicity, our nationality, our aesthetic. A neighbor’s suffering registers more viscerally than a stranger’s on the other side of the world, even when the stranger’s suffering is objectively more severe.

For day-to-day social bonds, this is manageable. Scaled to policy, it becomes a moral disaster. Immigration debates, foreign aid allocation, climate action — all of these involve weighing the interests of people who are distant, unfamiliar, often invisible. Empathy, left to its own devices, systematically discounts them.

What Works Better

The alternative Bloom proposes isn’t indifference. It’s rational compassion — a deliberate, considered form of caring that operates independently of emotional resonance. Where empathy asks “how does this make me feel?”, rational compassion asks “what action actually reduces suffering at scale?”

The distinction sounds cold. In practice, it’s what drives effective altruism, evidence-based charity evaluation, and the utilitarian frameworks behind public health policy. The people who have saved the most lives — Peter Singer’s philosophical tradition, the researchers behind deworming programs that improved outcomes for millions of African children — rarely operated on empathetic impulse. They operated on analysis.

Rational compassion can be moved by a child’s face. But it doesn’t stop there. It asks: how many children? Which intervention? What’s the cost per life saved? Empathy stops at the face. This bias is also explored from a different angle in small talk is the most sophisticated social skill—how the signals we read in brief exchanges often have nothing to do with the actual content of what’s being communicated.

The Uncomfortable Implication

None of this means empathy is worthless. In close relationships, in moments of personal grief, in one-on-one human contact, empathetic resonance is irreplaceable. The problem arises when we export it wholesale into domains — politics, justice, economics — where its biases cause structural harm.

The deeper discomfort is what this implies about moral seriousness. Genuinely ethical behavior, at scale, often requires overriding the emotional response rather than following it. Donating to the most effective cause rather than the most heart-wrenching one. Supporting policies that help the most people even when those people are abstract, distant, or unfamiliar.

Empathy feels like virtue. Rational compassion requires the harder thing: caring even when you can’t feel it.