Is your LLM Ageist?

ACL ARR 2026 January Submission3780 Authors

Keywords: LLMs; Age-related differences; Younger adults, Older adults; Prompting styles; LLM-as-a-judge

Abstract:

As society's reliance on LLMs becomes stronger with each passing day, the concerns over its treatment of less-tech-savvy and other discriminated demographics grow and grow.
We examine the effects of age on LLM use on both the input and output sides, over two tasks representing an assisted creativity scenario and an information request scenario.
The results of our human study suggest that LLMs do not significantly discriminate against younger or older adults, despite their clear stylistic differences. We do, however, find that assisting creativity is easier for models than providing information.
We also find that the practice of LLM-as-a-judge is a reliable proxy for self-evaluation questionnaires on these scenarios.

Paper Type: Short

Research Area: Computational Social Science, Cultural Analytics, and NLP for Social Good

Research Area Keywords: human behavior analysis; language/cultural bias analysis; human-computer interaction; sociolinguistics; NLP tools for social analysis

Contribution Types: Model analysis & interpretability, Data analysis

Languages Studied: English

Submission Number: 3780