HPV vaccination is central to cervical cancer prevention, yet uptake in China remains low, with information asymmetries and financial costs remaining major barriers. We study how subsidy design, procurement-price disclosure, and AI-assisted decision support shape HPV vaccine choice among 214 female undergraduates at Wuhan University. We randomly assign participants to six treatments that vary three subsidy schemes and two price transparency conditions. The subsidy schemes are baseline (a symbolic RMB 20 subsidy), targeted (a targeted RMB 247 per-dose subsidy applied only to the domestic bivalent vaccine), and universal (an RMB 247 per-dose subsidy applied to all vaccine types). The price transparency conditions vary whether participants are informed that the government procurement price of the domestic bivalent vaccine is below RMB 30. We also use a within-subject design in which participants interact with an AI health platform and then make their vaccination decisions before and after that interaction. We find that subsidies raise registration from about 70% to about 90%. AI, by contrast, leaves overall registration largely unchanged but shifts demand toward 9-valent vaccines, particularly the domestic 9-valent vaccine. Procurement-price disclosure reduces registration only in the baseline subsidy condition. Under targeted subsidy with procurement-price disclosure, the domestic bivalent share declines after AI exposure. These results offer practical guidance for policymakers and healthcare administrators designing vaccine promotion strategies under AI-assisted decision support.