Before I first moved to the United States at 13, my identity was simple: “I am from Saudi Arabia.” But as I got older, answering the question “Where are you from?” became increasingly complex. Because of school and family, I was constantly shifting between the two countries, never fully settling in one place.
In 2020, I started exploring my identity in hopes of an answer to this constant question. I interviewed my mother, tracing my ancestral history to the 1800s, and collected personal and archival imagery of my family and the regions they lived in. Through my mother’s beautiful retelling of our family’s story, I felt belonging.
As I processed the archival images through computer vision techniques — using a type of artificial intelligence that is designed to see and understand visual information as people do — I asked the A.I. follow-up questions about what it saw.
The answers revealed failures in the form of generalizations and stereotypes, uncovering how prejudice can be embedded in or learned by commercial A.I. tools. In the short film above, I converted the answers from the A.I. — results like “turban” or “dress” — into simple sentences, creating a character who my mother and I respond to with our own personal stories.
By juxtaposing oral storytelling with A.I.’s limitations, the film shows how artificial intelligence sees culture through a simplistic and biased lens, reducing my identity and erasing my ancestral stories and memories.
The post I Asked A.I. Where It Thought I Was
From. Its Answer: Nowhere. appeared first on New York Times.
