AI Mental Models & Trust: The Promises and Perils of Interaction Design

This study offers practical solutions to ongoing issues of trust and accountability in AI, highlighting how AI mental models are shaped among consumers in the evolving relationship between humans and AI. We argue that although predictability in AI is crucial, alone it is not enough to foster trust. The lack of real consequences for AI systems that breach trust remains a key challenge for interaction design. Until AI systems face tangible repercussions for trust violations, human trust will remain limited and conditional. Our research contributes to the development of socio‐technologies that prioritize human capabilities and foster productive human‐AI relationships.