Researchers find AI spontaneously develops a sense of facial beauty without training
New research shows that artificial neural networks can spontaneously develop a sense of facial beauty without any prior training. This discovery suggests that aesthetic preferences might stem from the basic ways our brains process visual information.
Researchers have discovered that artificial intelligence models can spontaneously develop a sense of facial beauty without any prior training or exposure to human preferences, according to a study published in Psychology of Aesthetics, Creativity, and the Arts. The research, conducted by Tianxin Shu and Delong Zhang from South China Normal University, suggests that this aesthetic sense emerges naturally from the way computer networks process basic visual information.
This finding adds to the ongoing debate about whether humans are born with an innate sense of beauty or if aesthetic preferences are learned through cultural exposure. The study involved generating a dataset of highly realistic human faces using an AI program and manipulating the underlying code to create varying levels of beauty.
These images were then fed into an untrained deep neural network called AlexNet, which has multiple layers of artificial neurons processing visual information. The researchers found that specific neural units in the network spontaneously began responding selectively to different levels of facial beauty, suggesting that this ability could be a byproduct of basic structural encoding rather than a learned behavior.
Brief written by urgent.news from PsyPost's own syndicated text. Machine-written — may contain errors; check the original before relying on it.
