Detecting Random-Like Human Behavior via Eye Tracking Using Deterministic Labeling

In many industrial environments, reliable decisionmaking is critical for ensuring safety and efficiency. However, factors such as fatigue, distraction, stress, and subjective bias can lead to random-like decision behavior, increasing the likelihood of errors and potentially severe consequences. To address these challenges, this paper introduces an eye-tracking-based approach combined with deep learning to detect randomness in decisionmaking through the analysis of gaze behavior. The proposed framework acts as a pre-filtering mechanism for identifying unreliable decisions in complex industrial tasks involving multiple choices.