As generative artificial intelligence (GAI) becomes increasingly prevalent, concerns regarding information disclosure have garnered considerable attention. In this research-in-progress paper, expand on existing theories of privacy disclosure by introducing eye tracking measures and an experiment which will investigate the effectiveness of an informational privacy nudging technique. We hypothesize that privacy nudges will moderate the relationship between privacy calculus and varieties of privacy fatigue. We introduce an expanded conceptual model and describe an eye tracking experiment that can validate it. This research will also offer insights into whether past models of privacy disclosure generalize in the new context of effective and safer GAI design.