Portrait of a Woman

Overview
Portrait of a Woman began as an experiment in how artificial intelligence visualizes women. After generating 1,000 images from the single prompt “portrait of a woman,” I noticed a striking uniformity—homogeneous expressions of beauty that flattened identity rather than expanded it. From there, I shifted focus from what a woman looks like to who a woman is.

Phase One — Composite Identity
I used the 2020 U.S. Census data and worked with ChatGPT to conceptually generate a portrait representing every woman in the United States—nearly 170 million individuals. Instead of producing an archive, I distilled the concept into a singular AI-generated composite—a layered abstraction that questioned the idea of collective identity rather than erasing individuality.

Phase Two — Story-Based Portraits
The project then embraced personal narrative. Through online surveys and outreach, I collected 250 stories of women who profoundly shaped someone’s life—mothers, mentors, friends, strangers. From these, I selected 50 accounts that balanced personal resonance with universal themes. Each story became the prompt for an AI-generated “story portrait”—a visual translation of memory, emotion, and influence. These are not literal likenesses but interpretive responses to who a woman becomes in someone else’s eyes.

Approach & Impact
Portrait of a Woman is a mosaic of 50 portraits, 50 stories, 50 ways of imagining womanhood beyond surface representation. By combining quantitative logic, personal narrative, and AI generation, the work highlights how identity is shaped not only by visible traits but by relationships, memory, and meaningful impact. The project invites viewers to look past the question of what a woman looks like and consider who she has been—and might still become.