Text-to-Image Generation of Depictions of Surgeons by DALL-E 2, Midjourney, and Stable Diffusion

Artificial intelligence (AI) image generation is becoming more widespread. There are potential pitfalls, such as perpetuating existing societal biases through learning and replicating biases. This study aimed to evaluate the demographic representation from 3 leading AI text-to-image generators. This video provides a demonstration of a user entering the prompt, 'a photo of the face of a surgeon,' into the command box of 3 text-to-image generator websites: DALLE-E 2, Midjourney, and Stable Diffusion. The output generated from each AI then demonstrates a variety of portraits of surgeons from different races and sexes to varying degrees, depending on the generator used. Click the Related Article link for full details.