Artificial generation of facial images is increasingly popular, with machine learning achieving photo-realistic results. Yet, there is a concern that the generated images might not fairly represent all demographic groups.
We use a state-of-the-art method to generate 10,000 facial images and find that the generated images are skewed towards young people, especially white women.
We provide recommendations to reduce demographic bias in artificial image generation.
Salminen, J., Jung, S.G., Chowdhury. S., and Jansen, B. J. (2020) Analyzing Demographic Bias in Artificially Generated Facial Pictures. ACM CHI Conference on Human Factors in Computing Systems (CHI’20) (Extended Abstract), Honolulu, HI, USA. 25–30 April, 1-8.