Expressing Social Emotions: Misalignment Between LLMs and Human Cultural Emotion Norms

TL;DR

This study evaluates how well large language models (LLMs) capture culturally specific social emotion expressions, revealing systematic misalignments and limitations in diversity compared to human behaviors across cultures.

Contribution

It introduces a psychologically informed evaluation framework and empirically assesses LLMs' ability to reflect cross-cultural social emotion patterns, highlighting significant gaps.

Findings

LLMs tend to overexpress engaging emotions compared to disengaging ones.

Responses from LLMs are highly concentrated and lack diversity in social emotion expression.

Model responses are robust to temperature changes but sensitive to prompt language and format.

Abstract

The expression of emotions that serve social purposes, such as asserting independence or fostering interdependence, is central to human interactions and varies systematically across cultures. As LLMs are increasingly used to simulate human behavior in culturally nuanced interactions, it is important to understand whether they faithfully capture human patterns of social emotion expression. When LLM responses are not culturally aligned, their utility is compromised -- particularly when users assume they are interacting with a culturally attuned interlocutor, and may act on advice that proves inappropriate in their cultural context. We present a psychologically informed evaluation framework of cross-cultural social emotion expression in LLMs. Using a human study comparing European American and Latin American participants' expression of engaging and disengaging emotions, we evaluate six frontier LLMs on their ability to reflect culturally differentiated patterns for expressing social emotions. We find systematic misalignment between model and human behavior: all models express engaging emotions more than disengaging ones, with particularly stark differences observed for the generally well-represented European American persona. We further highlight that LLM responses are highly concentrated and deterministic, failing to capture the diversity of human responses in expressing social emotions. Our ablation analyses reveal that these patterns are robust to sampling temperatures, partially sensitive to prompt language, and dependent on the response elicitation format. Together, our findings highlight limitations in how current LLMs represent the interaction of cultural and emotional axes, particularly when expressing social emotions, with direct implications for their deployment in cross-cultural affective contexts.