Virtual Mourning in the Metaverse: User Acceptance, Needs, and Limitations of AI-Based Memory Avatars

Abstract

Digital transformation is increasingly shaping the ways in which people remember and grieve. While existing platforms and social networks primarily preserve static content, immersive environments such as the metaverse and advances in generative artificial intelligence (AI) enable the creation of interactive, AI-based memory avatars, often also referred to as “griefbots” or “deadbots”. This study examines the perceptions, acceptance, opportunities, as well as psychosocial and ethical limitations of such avatars from the perspective of potential users. To provide an empirical foundation, an interactive prototype was developed and evaluated as part of a quantitative study (N = 33). These results were supplemented by qualitative guided interviews with experts in grief counseling, psychotherapy, and pastoral care. The findings paint a highly ambivalent picture: While respondents see potential in preserving life stories or in processing missed goodbyes, concerns regarding emotional dependence, difficulties in accepting loss, and ethical and data protection risks dominate. The Behavioral Intention to use the technology depends largely on Perceived Usefulness (r = 0.74) and Emotional Impact (r = 0.82), rather than purely on user interface elements. Experts emphasize that simulated interactions must not hinder the necessary process of letting go. The study concludes with recommendations for transparency requirements, time limits, and clear ethical safeguards for vulnerable groups.