AuthZit: Personalized Visual-Spatial and Loci-Tagging Fallback Authentication

Designing a fallback authentication that is both memorable and strong poses a challenging task due to the need for authentication secrets to remain secure and easily recallable without frequent reinforcement. This could be especially prevalent for cloud computing security and resiliency. Inspired by the robust visual-spatial memory and associative memory of individuals, we introduce AuthZit, a novel system. AuthZit encodes authentication secrets as paths implementing a fault-tolerant algorithm through a 3D map of real-life places, navigated in both first person and 2D bird’s-eye perspective, coupled with a loci-tag (textual secret) associated with the location. Two experiments were conducted to iteratively design and evaluate AuthZit. First, it was observed that visual-spatial secrets are most memorable when navigated through a combination of 3D first-person and 2D bird’s-eye view perspectives. Second, we evaluated AuthZit against security questions and Android’s 9-dot pattern lock across three dimensions: memorability, security, and speed. AuthZit’s complexity-controlled secrets were significantly more memorable after three months, more resilient to shoulder surfing, and close adversaries.