Irony in Emojis: A Comparative Study of Human and LLM Interpretation

TL;DR

This paper investigates GPT-4o's ability to interpret irony in emojis, comparing its performance with human perceptions and analyzing the influence of demographic factors on interpretation accuracy.

Contribution

It provides a comparative analysis of GPT-4o and human understanding of ironic emojis, highlighting interpretive strengths and limitations of LLMs.

Findings

GPT-4o shows partial alignment with human perceptions of irony in emojis.

Demographic factors like age and gender influence emoji interpretation.

GPT-4o's performance varies across different demographic groups.

Abstract

Emojis have become a universal language in online communication, often carrying nuanced and context-dependent meanings. Among these, irony poses a significant challenge for Large Language Models (LLMs) due to its inherent incongruity between appearance and intent. This study examines the ability of GPT-4o to interpret irony in emojis. By prompting GPT-4o to evaluate the likelihood of specific emojis being used to express irony on social media and comparing its interpretations with human perceptions, we aim to bridge the gap between machine and human understanding. Our findings reveal nuanced insights into GPT-4o's interpretive capabilities, highlighting areas of alignment with and divergence from human behavior. Additionally, this research underscores the importance of demographic factors, such as age and gender, in shaping emoji interpretation and evaluates how these factors influence GPT-4o's performance.