This Thesis entitled Artificial Interaction: Analyzing The Case of Virtual Influencer @lav_caca Towards User's Perception of Dangdut Music on Instagram, aims to understand the phenomena of AI-generated influencers on social media, specifically on Instagram platform. The paradigms and methods used in this research are the paradigm of phenomenology with qualitative research methods, and case study analysis. The results of this study found that Instagram users develop one-sided relationships with virtual influencer Laverda Salsabila, contrary to those formed with traditional human celebrities. Attributes of Caca as a "metahuman" Dangdut influencer also affects users' perceptions, such as authenticity, relatability, and aesthetic appeal. Through interview sessions with informants, and content analyses of Instagram account @lav_caca indicates that novelty and humanization significantly affect user engagements with virtual influencers, particularly among Gen-Z as the younger audiences. The study of AI-generated influencers in social media marketing may provide some practical benefits for marketers and practitioners. By understanding the consumers' perceptions, limitations, and challenges of virtual influencers in social media, brands can make informed decisions about adopting and implementing AI technologies to improve their marketing strategies and gain a competitive advantage; since, innovation and adaptation are the future of marketing.

Keyword: virtual influencers, Instagram, social media, artificial intelligence