2026
INVISIBLE TO THE MACHINE.
An Interactive Installation Exploring Bias in Facial Recognition Systems
TOOL TouchDesigner, Arduino, Adobe After Effects VIDEO Video Link: https://youtu.be/kr0HFF8uv7I This interactive installation explores how facial recognition systems respond differently to faces under changing conditions. The project focuses on skin tone and lighting as two factors that can affect recognition. A camera captures the participant’s face while the system tracks facial features in real time. Changes in light can affect the detection result. When recognition becomes less stable, parts of the face are incorrectly identified or replaced. These errors appear directly on the screens as visual distortions. By making recognition errors visible, the installation gives participants a direct view of how a machine reads a human face. It questions the idea that facial recognition is equally accurate for everyone.