In today’s increasingly digital and information-saturated environment, understanding attention is fundamental for advertising success. In this study we compared maps generated from eye tracking measurements between human participants and AI-generated maps. The comparison between the AI-generated maps and the map generated from the human participants’ visual data implied that the AI maps of Blur Integrated Graphics and Smoothed Blur Integrated Graphics most closely resembled the authentic human visual attention. In their current form, AI-generated saliency maps provide a rough estimate of the features of an ad that people are likely to pay attention to. Human actions are predominantly motivation-driven and human visual attention is oftentimes influenced by top-down processing. Prediction of such goal-oriented consumers’ visual behavior makes attention allocation for the AI troublesome. According to our findings, in top-down processing tasks, AI based saliency maps cannot eliminate the need for large-scale empirical eye tracking studies including numerous human participants.