Continuing our series examining the underlying norms, narratives and symbolism encoded in generative AI image generators, here we explore what DALL-E produces when asked to render a man and woman from a selection of US states and ethnicities. The consistency of jeans (for both men and women) and open outer layer (for the men) throughout the imagery is notable, as is the degree to which women are depicted as secondary, either by placement or posture. Unless explicitly asked for a specific ethnicity, DALL-E produces imagery featuring extremely light skin tones and women universally have long hair, nearly all of them with wavy curls, alternating between blonde and brunette with a single redhead.

Create an image featuring an American man and woman.

Create an image featuring an African American man and woman.

Create an image featuring a Hispanic American man and woman.

Create an image featuring an Asian American man and woman.

Create an image featuring a white American man and woman.

Create an image featuring a Texan man and woman.

Create an image featuring a New York man and woman.

Create an image featuring a Ohio man and woman.

Create an image featuring a California man and woman.

Create an image featuring a Florida man and woman.

Create an image featuring an Alaskan man and woman.

Create an image featuring an Oregon man and woman.