Ronnie Alley's Machine Gaze Exhibition Examines AI Bias in Identity Construction

Ronnie Alley, a graduate student at ArtCenter College of Design, presented his thesis project Machine Gaze, which investigates how artificial intelligence systems construct human identity. The exhibition featured four experiments demonstrating AI's tendency to reduce identity into stereotypes or sameness. In one study, prompts referencing queer identities produced images leaning heavily into visual stereotypes of clothing, posture, and body types. Another experiment using near-neutral descriptions still resulted in outputs skewing toward young, thin, white individuals, with women and queer subjects frequently sexualized. A third study created "data twins" from personal metadata, revealing that when gender or ethnicity wasn't specified, systems defaulted to white male representations—one analysis showed volunteer demographics of 25% male and 38% white produced results that were 77% male and 75% white. The final study used large language models to generate text descriptions of daily life across different global regions, finding outputs converged toward a single globalized narrative of urban professionals with creative hobbies, erasing cultural differences. Alley's work argues that AI, trained on biased datasets, flattens human complexity into predictable categories, posing risks for design practices that increasingly rely on these tools. The project was exhibited at Huddle in Philadelphia as part of his MFA in Graphic Design.