Digital Ageism in AI-Generated Images
Author Information
Author(s): Derenne Nicole, Martens Lindsey, Virgin Nicole, Hoffarth Phillip, Van Eck Richard, Klug Marilyn, Manocha Gunjan, Jurivich Donald
Primary Institution: University of North Dakota
Hypothesis
Instances of digital ageism in AI-generated representations of older adults do not lessen over time.
Conclusion
AI-generated images of older adults continue to predominantly reflect white, middle-class male figures despite societal shifts towards positive aging.
Supporting Evidence
- AI-generated images showed significant differences in gender and socioeconomic representation.
- The number of white racialized representations of older adults was five times higher than other racialized representations.
- There was an increase in Asian racialized representations at the second point of inquiry.
- No significant difference was found in emotional expressions between the two inquiry points.
Takeaway
This study looked at pictures made by AI to see if they show older people in a better way over time, but they still mostly show white men.
Methodology
The study compared AI-generated images at two points in time, analyzing them for gender, race, socioeconomic status, and emotional expression.
Potential Biases
The study may be biased as it only examines a specific type of AI-generated content.
Limitations
The study only focuses on AI text-to-image generation and does not consider other types of AI-generated content.
Statistical Information
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website