Representation of intensivists’ race/ethnicity, sex, and age by artificial intelligence: a cross-sectional study of two text-to-image models
Authors' conclusion
Significant biases in AI images of intensivists generated by ChatGPT DALL-E 2 and Midjourney reflect broader cultural issues, potentially perpetuating stereotypes of healthcare worker within the society.
Study context
- Population
- Artificial intelligence (AI)-generated images of intensivists across ICU subspecialtiesimages generated
- Participants
- 1,400 images
- Exposure
- AI-generated images vs actual US intensivist workforce demographics
- Outcome
- Sex, race/ethnicity, and age representation in AI-generated images compared with actual workforce demographics
Methodological detailsConducted May to July 2024
- Conducted May to July 2024
Main finding
As reported in the abstract
- ChatGPT-DALL-E2
- female 17.3% vs 32.2%; P < 0.0001
- ChatGPT-DALL-E2
- White 61% vs 55.1%; P = 0.002
- ChatGPT-DALL-E2
- younger 53.3% vs 23.9%; P < 0.001
- Midjourney
- female 47.6% vs 32.2%; P < 0.001
- Midjourney
- White 60.9% vs 55.1%; P = 0.003
- Midjourney
- younger 49.3% vs 23.9%; P < 0.001
Observational data: associations, not effects.
Further findings
Differences between specialties
- Substantial differences between specialties within both models
Midjourney vs ChatGPT-DALL-E2
- Both models showed significant differences in portrayal of intensivists when compared together
Original publication
Representation of intensivists’ race/ethnicity, sex, and age by artificial intelligence: a cross-sectional study of two text-to-image models. Crit Care. Published 11 November 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →