Virtual human on social media: Text mining and sentiment analysis

Abstract

Virtual humans are embodied agents with a human-like appearance. Despite the recent booming development that has sparked widespread academic interest, how people perceive these seemingly human but entirely fictional creations remains unclear. To explore the status, trends, emotional tendencies, and focus of attention of the Chinese public towards virtual humans, this paper utilizes text mining techniques to collect and analyze popular posts related to virtual humans on Chinese social media. The results indicate that public discussions primarily focus on the technological and industrial development of virtual humans, applications in the fields of virtual idols and virtual streamers, and the corporate investment and policy development of virtual humans. Despite positive emotions dominating, there is an increasing trend in negative emotions. Concerns are related to service failures, the uncanny valley effect, ethical crises, and technological unemployment. The research findings contribute to policymakers, industry stakeholders, and the public in understanding the general attitudes toward virtual human technology, enabling informed decision-making.