How AI digital-human livestreaming influences consumers’ purchase intentions: evidence from Luo Yonghao’s “6.18” livestream in 2025

Abstract

Generative artificial intelligence (AI) is developing rapidly, and AI digital human livestreaming has become a key operational model for e-commerce platforms. This study, using Luo Yonghao's AI digital human livestream during the 2025 "6.18" shopping festival as an example, constructs a four-dimensional framework from the consumer perspective: "Source credibility - Cognitive fluency - Parasocial-social interaction - Perceived novelty." The research is conducted through social media content analysis and semi-structured interviews. Results indicate that Source credibility is the most prominent (35%), followed by cognitive fluency (26%), parasocial-social interaction (19%), and perceived novelty (15%). Trust can be smoothly transferred from the existing persona to the digital avatar. Clear, structured explanations reduce comprehension costs. Familiarity and a sense of companionship enhance presence and promote conversions. Novelty primarily contributes initially and decays quickly. Based on this, it is recommended to leverage established IP for trust transfer, combine "product interaction/social interaction" by category, and strengthen long-term trust through human-machine co-streaming and transparent disclosure. This study aims to summarize and validate the key pathways by which AI digital human livestreaming influences consumer purchase intention and provide a reference for the design and management of livestreaming for platforms and brands.