Designing LED Expressions in An Interactive AI Product

Abstract

As AI is increasingly integrated into everyday products, LED expressions offer a compact and intuitive channel for communication. While LEDs are widely used as simple indicators, their expressive potential remains underexplored in design research. This study aims to investigate how LEDs can be designed to convey various types of messages more expressively and communicatively. To this end, we adopted a three-step research process based on a design methodology: (1) design ideation workshop to collect possible messages and their corresponding expression, (2) creating LED expression GIF animations following the design guidelines, and (3) an online survey to evaluate the appropriateness of each LED expression-message pair. The survey results were analyzed using both mean scores and coefficient of variation (CV), which allowed us to classify the LED expressions into four categories based on interpretability and consistency: universal & accurate, competitive & contextual, abstract & diverse, and inappropriate. Our findings show that users intuitively map specific message characteristics—such as arousal, urgency, or process—to LED design factors, such as intensity and rhythm. This highlights the potential of LED expressions as a viable communication channel for an AI product. Future research should address the limitations of the study, including the absence of physical LED testing, the exclusion of color as a design factor, and the lack of consideration of cultural diversity.