The rapid growth of short-form video platforms such as TikTok, Instagram Reels, and YouTube Shorts has reshaped digital communication and marketing, making the characteristics of on-screen presenters strategically significant. Meanwhile, advances in artificial intelligence have enabled AI-generated speakers, synthetic presenters delivering scripted content through virtual avatars and AI-driven voice systems, which brands increasingly adopt for their scalability and flexibility. Despite their growing prevalence, it remains unclear whether AI-generated speakers can match the persuasive effectiveness of human communicators. The Elaboration Likelihood Model (ELM) explains persuasion through cognitive evaluations of argument quality (the central route) and heuristic assessments of source credibility (the peripheral route). However, these models were developed in human-centric contexts, leaving it uncertain whether established persuasion mechanisms remain stable when the communicator is technologically mediated. Recent work suggests that disclosing AI authorship can lower message evaluation, yet in short-form video settings characterized by high informational density and rapid processing, persuasion may depend more on content quality than on the speaker's ontological status.

JEL Codes:

Keywords: AI-generated speakers, elaboration likelihood model, short-form video persuasion, perceived authenticity