משפיענים ברשת: ניתוח הבדלים באינטרקציות עם משפיענים וירטואליים לעומת משפיענים אנושיים

AI applications are increasingly integrated into social networks. In this project, we investigated the differences between Instagram users' interactions with virtual influencers vs. those with human influencers. We found that virtual influencers receive significantly less attention than human influencers. This is reflected in fewer likes, more messages with neutral sentiment, and shorter responses compared to interactions with human influencers. Additionally, we developed a machine learning model that uses a set of parameters to measure user interactions, allowing us to classify Instagram influencers as human or virtual (F1-score 0.89). This study deepens our understanding of the role of virtual influencers in social networks and enables the improvement of marketing strategies that utilize virtual influencers.