Text-to-image diffusion models are increasingly deployed in creative contexts, yet remain impersonal — optimized for aggregate …