AGE-RELATED VALUE ORIENTATIONS IN LARGE LANGUAGE MODELS (LLMS)
2024

Age-Related Value Orientations in Large Language Models

Sample size: 80 publication Evidence: moderate

Author Information

Author(s): Zhang Xin, Ren Yuanyi, Song Guojie

Primary Institution: Peking University

Hypothesis

How do Large Language Models (LLMs) exhibit age-related value orientations compared to human participants?

Conclusion

Large Language Models showed less hostile ageism and a similar level of benevolent ageism compared to humans, while also exhibiting a high-warmth, high-competence stereotype towards older adults.

Supporting Evidence

  • LLMs exhibited significantly less hostile ageism compared to human participants.
  • LLMs showed a similar level of benevolent ageism to that of human participants.
  • LLMs displayed a High-Warmth-High-Competence stereotype towards older adults.

Takeaway

This study looked at how AI models think about older people and found that they are nicer than humans in some ways.

Methodology

The study used a newly developed evaluation pipeline to assess the age-related value orientations of six different LLMs and conducted an experiment with younger adults discussing age-related topics with LLMs or humans.

Participant Demographics

80 younger adults were randomly assigned to the study.

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1093/geroni/igae098.3251

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication