Article | Open Access
Personalized Persuasion Through Conversational AI: Can DeepSeek Change Perceptions of Genetically Modified Foods in China?
| Views: | 38 | | | Downloads: | 12 |
Abstract: Conversational AI has become an influential intermediary in public communication. Emerging research on conversational AI highlights its potential to correct misconceptions and influence attitudes across domains. This study investigates the persuasive effects of personalized conversational AI, focusing on genetically modified foods in China. Employing a between-subjects factorial design, 813 participants engaged in dialogues with a DeepSeek-based chatbot. Participants were randomly assigned to one of four conditions, ranging from a non-personalized generic control to increasingly tailored approaches based on demographic information, risk perceptions, or a combination of both. Results indicate that while AI interactions significantly improve attitudes and willingness to consume genetically modified foods across all conditions, the additional persuasive effect of personalization was conditional. Only a personalization strategy combining demographics and risk perception yielded greater persuasive effects than the control, primarily among participants with positive risk perceptions. Furthermore, moderation analyses revealed a divergence in individual differences: Among participants with negative risk perceptions, while greater prior experience with AI and higher trust in science decreased the persuasive effects, higher AI knowledge facilitated greater attitude gains.
Keywords: AI persuasion; conversational AI; generative AI; genetically modified foods; science communication
Published:
Issue:
Vol 14 (2026): Exploring Engagement With Complex Information: Perspectives on Generative AI as an Information Intermediary (In Progress)
© Qi Xi, Jing Zeng, Zhanghao Li, Mike S. Schäfer. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 license (http://creativecommons.org/licenses/by/4.0), which permits any use, distribution, and reproduction of the work without further permission provided the original author(s) and source are credited.