Generating Visual Aids to Help Students Understand Graphic Design with EKPHRASIS
‍Bob Tianqi Wei and Shayne Shen, Shm Almeda, Bjoern Hartmann
CHI 2025, LBW PAPER UNDER REVIEW
‍
ABSTRACT
Graphic design relies on complex language that creates significant barriers in design education, especially for students from non-design backgrounds or those with language barriers. Our interview with 11 design educators and professional designers confirmed that this specialized language—characterized by jargon, ambiguous terminology, and underspecified words—operates within insular communities, making it challenging for learners to decode and internalize its meaning. To address this challenge, we developed EKPHRASIS, an interactive educational system powered by a machine-learning model that provides real-time and actionable visual aids to support students' subjective understanding of design language. Preliminary user testing with novice participants suggests that EKPHRASIS enhances intuitive, multi-modal learning and teaching in graphic design. Through EKPHRASIS, we explore the potential for AI technologies to democratize design pedagogy with real-time, context-aware multi-modal learning support.