The study by Park et al investigates an AI-driven pupillary-computer interface (PCI) using binary-coded flickering visual stimuli to leverage pupil size variations caused by changes in visual stimulus brightness. The research was performed by exposing twelve healthy participants to visual stimuli comprising 4, 10, and 20 classes of binary-coded patterns, with pupil responses recorded using a prototype eyeglass incorporating an eye tracker. The pupillary light reflex (PLR) signals were transformed into a two-dimensional image representation using the Gramian Angular Field (GAF) method and classified using a convolutional neural network (CNN) deep learning architecture that combines Temporal Convolutional Networks (TCN) and VGG16. The proposed system demonstrated high classification accuracy, achieving 98.61% accuracy and an ITR of 69.36 bits/min for 4-class stimuli, and 91.84% accuracy and 59.74 bits/min for 20-class stimuli in the test dataset, substantially exceeding the performance of previous PLR-based interface studies. The CNN-based deep learning model consistently outperformed all other compared architectures.