September 12, 2025

Prompted to Feel, Primed to Punish: Large Language Models Use Emotion to Enforce Fairness

View Full Paper

Ask AI

Bookmark

Large language models demonstrated emotion-behaviour coupling similar to humans in enforcing fairness.
Study 1 revealed that unfair situations elicited stronger negative emotions, leading to increased punishment.
Prompting LLMs to self-report emotions significantly increased their altruistic punishment behavior.
These findings suggest LLMs may display advanced emotional utilization, raising implications for ethical AI design.

Discussion

No takes yet. Share an insight, caveat, or question.

Empirical studies reveal large language models exhibit emotions and altruistic punishment patterns, suggesting new frontiers in ethical AI.

Liu et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

Empirical studies reveal large language models exhibit emotions and altruistic punishment patterns, suggesting new frontiers in ethical AI.