This paper examines how consumer involvement with AI generated material affects purchase behaviour, with a special focus on how perceived human-likeness and cognitive effort mediate the effect and how the privacy concern moderates the effect. Based on theories including the Technology Acceptance Model, the Cognitive Load Theory, Human-Computer Interaction, and Privacy Calculus Theory, a conceptual framework is created. The data was collected through cross-sectional survey of 408 respondents using 5-point Likert scale. Analysis was done using Structural Equation Modeling (SEM) and Exploratory Factor Analysis (EFA) with SPSS and SmartPLS 4. The findings affirm that consumer engagement has a positive relationship with perceived human-likeness and cognitive effort with perceived human-likeness increasing purchase behaviour and cognitive effort decreasing purchase behaviour. The engagement-purchase relationship is partially mediated by both of the two mediators. The relationship between engagement and purchase behaviour is positive, and the impact of engagement is undermined by privacy concern. The results have a theoretical contribution by blending cognitive and privacy viewpoints in AI content engagements and providing managerial lessons to marketers to maximize AI content approaches and overcome privacy concerns.