Researchers found that while women are increasingly represented as doctors in AI-generated narratives, they predominantly embody traditional roles, like nurses, perpetuating stereotypes.
According to the study, 98% of nurses generated were identified as women, while women made up 50% to 84% of stories about doctors, highlighting algorithmic bias.
Dr. Sarah Saxena points out that despite attempts to correct biases in AI, the current data may reflect an overcorrection in gender representation in medical roles.
The research indicates that AI models have integrated social biases from their training data, where traditional gender roles still dominate the perception of medical professionals.
[
Collection
]
[
|
...
]