When Large Language Models are Reliable for Judging Empathic Communication
This work addresses the problem of ensuring reliable oversight in emotionally sensitive AI applications, such as conversational companions, by providing a validation method for LLMs, though it is incremental in benchmarking performance.
The study investigated the reliability of large language models (LLMs) in judging empathic communication by comparing annotations from experts, crowdworkers, and LLMs across four evaluative frameworks applied to 200 real-world conversations, finding that LLMs consistently approach expert-level reliability and exceed crowdworker reliability.
Large language models (LLMs) excel at generating empathic responses in text-based conversations. But, how reliably do they judge the nuances of empathic communication? We investigate this question by comparing how experts, crowdworkers, and LLMs annotate empathic communication across four evaluative frameworks drawn from psychology, natural language processing, and communications applied to 200 real-world conversations where one speaker shares a personal problem and the other offers support. Drawing on 3,150 expert annotations, 2,844 crowd annotations, and 3,150 LLM annotations, we assess inter-rater reliability between these three annotator groups. We find that expert agreement is high but varies across the frameworks' sub-components depending on their clarity, complexity, and subjectivity. We show that expert agreement offers a more informative benchmark for contextualizing LLM performance than standard classification metrics. Across all four frameworks, LLMs consistently approach this expert level benchmark and exceed the reliability of crowdworkers. These results demonstrate how LLMs, when validated on specific tasks with appropriate benchmarks, can support transparency and oversight in emotionally sensitive applications including their use as conversational companions.