#517: Games of Representation: Making the Abstract Tangible through Data Representation with Card-Based Games for Understanding AI Ethics Education

Middle school students notice that an image search for "physicist" returns predominantly white male faces, but when asked "Who is represented? Who is left out?" they struggle to respond. This reveals a critical gap in AI literacy education: youth lack understanding of representation and the complex interchange between societal patterns and AI systems. Games of Representation (GR) addresses this gap through card-based games that make abstract concepts of dataset representation tangible. Aligned with the Kapor Foundation's Responsible AI and Tech Justice Guide, the design engages players as stakeholders who curate datasets using SET cards. This paper contributes two design principles: (1) scaffolding from individual to system-level understanding of how bias emerges in data, (2) connecting mathematical representation (statistical distributions) to social representation (who gets included or excluded). These principles offer guidance for developing educational responses to emerging technologies where technical and social dimensions are inseparable.