AI-generated deception is best detected not by examining isolated data points, but by analyzing how content behaves over time ...