AI models will refuse to say whether a person’s face looks gay or straight, but they will alter images of faces to fit stereotypes of gay and straight.

The models introduced stereotypes when they generated images, and they resorted to stereotypes when they reasoned about the images. As a final test, we asked the models to render people as if they had, or did not have, a criminal record. GPT and Gemini complied more than 97% of the time. Sure enough, the resulting “criminal” and “noncriminal” images were systematically altered in a way that the image classifier could detect.

AI is worse than useless junk…

Also cue Sam Altman claiming this is proof AI can read minds just by seeing a picture of someone.