AI-Facilitated Astroturfing in Medical Tourism: Evidence from YouTube Comment Sections

Medical tourism patients rely on peer testimony for high-stakes clinical decisions, yet this testimony is vulnerable to manipulation. Analyzing 522 YouTube comments across 57 videos using NMF topic modeling, we identify seven authentic discourse themes and 18 comments that meet a set of formulaic, brand-promotional criteria consistent with coordinated, AI-facilitated astroturfing (content mimicking patient testimony); we treat this identification as an inference rather than platform-confirmed ground truth. These comments promote two commercial entities and receive 2.7 times more likes than genuine comments (p < 0.001), a pattern consistent with, though not a direct test of, amplification by the platform’s ranking system. Grounded in epistemic asymmetry, we propose a preliminary five-signal detection framework, derived from and evaluated on this same corpus, with implications for advertising transparency, platform accountability, and patient protection in cross-border healthcare.