A Visual Audit of Race, Class and Family Norms in GenA
Created by: Qianhe Gong, Yichun Huang, Jiayan Ma, Wenxuan Zhang
Overview: This project uses prompt-based visual auditing across three AI image generators to expose how GenAI constructs a hierarchy of family visibility—normalizing white, heterosexual, middle-class nuclear families as the unmarked norm while stereotyping, correcting, or conditionally including non-Western and non-normative family structures.
Created with: DouBao, Gemini, ChatGPT
Concept
Poster description: This poster presents a prompt-based visual audit of three AI image generators, examining how “happy family” is algorithmically constructed across race, region, and family structure. It reveals a visibility hierarchy where nuclear middle-class families appear as the unmarked norm, while non-dominant families are stereotyped, corrected, or conditionally included.
Development and Sketches
Contact the Creators
(Include group project members contact details, email, social media accounts and links here if desired )