The personalisation paradox: How generative AI shapes conservation intentions through emotional and cognitive pathways

Ocean ecosystems face unprecedented threats, yet public engagement in this area remains relatively low. Can generative artificial intelligence (AI) bridge this awareness-action gap? This study investigated how users respond to AI-powered ocean conservation education by examining three technological affordances (i.e., authenticity, knowledge acquisition, and personalisation) and their impact on conservation intentions. Using structural equation modelling on a dataset collected from 469 participants who experienced an AI conservation platform simulation in Singapore, we tested how these affordances influence emotional resonance and cognitive engagement, which subsequently shape self-efficacy and behavioural intentions. Contrary to expectations, the results revealed that personalisation was counterproductive, significantly reducing emotional resonance and cognitive engagement. Instead, users responded most strongly to authentic content and comprehensive knowledge delivery. These affordances triggered both emotional resonance and cognitive engagement, which together drove self-efficacy and behavioural intentions. The findings revealed a “personalisation paradox” in AI-mediated environmental education: while users expect customisation, they emotionally connect more with universal, authentic narratives about ocean conservation. This challenges prevailing assumptions about AI design and suggests that environmental educators should prioritise content credibility and knowledge depth over algorithmic personalisation. Theoretically, this study extends affordance theory by identifying boundary conditions where typical technological benefits reverse, offering practical guidance for designing AI-mediated environmental education that effectively motivates conservation action.