AI generates well-liked but templatic empathic responses

ArXi:2604.08479v1 Announce Type: new Recent research shows that greater numbers of people are turning to Large Language Models (LLMs) for emotional, and that people rate LLM responses as 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.