Artificial intelligence models find human faces more attractive than humans do, according to a new study. The research suggests AI systems tend to rate all faces as attractive, regardless of how they actually look.
AI rates all faces as attractive
The finding was revealed in a study that asked numerous AI models to compare the attractiveness ratings of 2,513 human faces. Researchers showed the faces to Claude, Gemini, ChatGPT and Grok to see how they rated human attractiveness and to find out if they had any biases.
Study author Macken Murphy, an expert in human mating behaviour, said AI models make humans feel attractive “regardless” of how they actually look. He told the Daily Star: “AI models overrate human attractiveness. They have a tendency to call faces attractive regardless of how they look.”
AI agrees with humans on ranking
Researchers said the bots rated faces in a very narrow range, avoiding very low and very high scores. Despite considering everyone as attractive, the experts said the bots could definitely tell whether one face is more or less attractive than another. This is because the machines agreed with humans on the relative attractiveness of faces.
Fellow study author Santiago Grandas said: “Humans and AI seem to agree in the rank-ordering of faces. If we asked a group of people to order a set of faces from least to most attractive, we would expect an AI model to roughly reproduce that same order.”
AI prefers women and younger faces
Grandas, the head of research at personal beauty consultancy Qoves, said AI took a particular liking to one group of people – much like humans do. He said: “On average, AI models rated women’s faces higher than men’s faces and also judged younger people as more attractive. We know from previous research that this is also the case for human raters, so there seem to be some facial characteristics that impact both human and AI evaluations of attractiveness.”
The study, published in Research Gate, delves into the emerging field of “machine psychology”. The development marks a stark change from the various warnings that AI could be developing too quickly and may have catastrophic consequences for mankind.
Contrast with AI warnings
It follows warnings that AI could wipe out humanity. Tech billionaire Bill Gates has warned AI could wipe out a billion humans in a Terminator-style apocalypse. He claimed there’s “never been a weapon as powerful … as people with ill intent using AI tools”.
The Microsoft co-founder has called for greater government oversight of the emerging tech, urging the US to take the lead in establishing safeguards.
