AI generates well-liked but templatic empathic responses

2026-04-09

AI Summary

This study investigates why empathetic responses generated by large language models (LLMs) are often preferred by users over those written by humans, uncovering the underlying linguistic patterns and limitations. By developing a taxonomy encompassing ten empathy strategies and combining manual annotation with quantitative analysis, the authors systematically examine the discourse structures of 3,265 AI-generated and 1,290 human-written empathetic responses. The research reveals, for the first time, that LLMs heavily rely on fixed templates: 83%–90% of AI responses follow an identical sequence of structured strategies, covering 81%–92% of their content—a significantly lower diversity compared to human responses. This work introduces the first function-oriented, structural analytical framework for empathetic discourse and quantifies the tension between the efficiency and template-driven nature of AI-generated empathy.

Abstract

Recent research shows that greater numbers of people are turning to Large Language Models (LLMs) for emotional support, and that people rate LLM responses as more empathic than human-written responses. We suggest a reason for this success: LLMs have learned and consistently deploy a well-liked template for expressing empathy. We develop a taxonomy of 10 empathic language "tactics" that include validating someone's feelings and paraphrasing, and apply this taxonomy to characterize the language that people and LLMs produce when writing empathic responses. Across a set of 2 studies comparing a total of n = 3,265 AI-generated (by six models) and n = 1,290 human-written responses, we find that LLM responses are highly formulaic at a discourse functional level. We discovered a template -- a structured sequence of tactics -- that matches between 83--90% of LLM responses (and 60--83\% in a held out sample), and when those are matched, covers 81--92% of the response. By contrast, human-written responses are more diverse. We end with a discussion of implications for the future of AI-generated empathy.

Problem

• empathy
• Large Language Models
• templatic responses
• empathic language
• AI-generated empathy

Innovation

• empathic template
• large language models
• discourse tactics
• formulaic responses
• AI empathy