AI-Native Persuasion Strategies for Structured Persuasion Message Design
Yu Chen, GuoJie Ma, Xiangling Zhuang · International Journal of Human-Computer Interaction · 2026
To enhance the persuasiveness of Artificial Intelligence (AI), previous research focused on adapting human persuasion strategies to AI, rather than leveraging AI-native strengths. We developed two strategies utilizing AI’s sensory sensitivity and data processing capabilities as supporting evidence within a tripartite (target + reason + support) message framework.In a set meal design task (N = 360), participants were persuaded to change food choices across four strategies (target-only, target-reason, and two AI-native support strategies) and two agent types (robot, virtual agent). Results showed that adding AI-native supports significantly increasedbehavioral compliance and perceived persuasiveness, while standard target-reason strategy yielded no such gains.Mediation analysis indicated that increased compliance was primarily driven by perceived persuasiveness, rather than perceived message credibility or trust in the agent. These findings suggest that effective AI persuasion can stem from leveraging machine-unique strengths without emulating human social dynamics, providing novel guidance for designing persuasive AI systems.