Abstract This article introduces and analyzes the notion of aesthetic bias in generative AI. I argue that AI models systematically privilege certain aesthetically relevant features while marginalizing others. Drawing on examples from popular text-to-image generators, I highlight recurring tendencies, such as a preference for beauty, spectacle, saturation, or symmetry. These aesthetic biases carry specific risks: they disguise aesthetic preferences as neutrality, threaten to homogenize artistic production and taste, and contribute to creating self-enclosed communities of appreciators, or “aesthetic bubbles.” The upshot is that aesthetic bias in AI demands serious study.
Alexandre Declos (Sat,) studied this question.