Study Finds AI Empathy Rated Higher Than Human

Penn State researchers report that people judged AI-written empathetic responses as higher quality yet still preferred human empathy.

Updated on Nov. 2, 2026 in Artificial Intelligence

Study Finds AI Empathy Rated Higher Than Human

Live Poll

Would you rather receive emotional support from a human than from an AI?

Penn State researchers published a study in the journal Communications Psychology finding that human users rated AI-generated empathetic responses higher in quality than human-generated ones, though they preferred human empathy overall. The researchers termed the effect the AI empathy choice paradox.

Why it matters

The findings suggest that perceived effort, rather than response quality alone, shapes how people value empathy. For Pennsylvania readers, the study bears on how AI tools used in the state's hospitals, schools, and customer services may be perceived.

Participants rated each response on perceived empathy and perceived effort, and perceived effort correlated with perceived empathy. The extraction does not report sample sizes or effect sizes, leaving the strength of the preference gap unquantified.

The players

Anat Perry

A researcher whose paper "AI will never convey the essence of human empathy" argues that human empathy involves effortful emotional exertion that AI cannot replicate.

The details

Participants evaluated each response for empathy and perceived effort, and the researchers observed that AI-generated responses were rated as exerting more effort than human ones. The study built on Anat Perry's argument that human empathy involves effortful emotional exertion, which AI does not experience.

- Article about the AI empathy study published November 2, 2026.

The Tech Race

The study extends Anat Perry's theoretical argument that effortful emotional exertion underpins human empathy by testing it experimentally against AI-generated responses. It adds evidence to the debate over whether conversational AI can substitute for human emotional support in contexts such as healthcare and counseling.

Users of AI chat and support tools may find AI responses sound more caring yet still prefer human contact when emotional support matters. The findings suggest that people may discount AI empathy even when its written quality is rated highly.

The takeaway

Perceived effort shapes how much people value empathy, so AI responses may need to signal effort to be trusted. Readers should weigh the quality of an empathetic message alongside who or what delivers it.