Productive Failure-based AI Literacy Escape Room (ProFAILER): Exploring Productive Failure and Gamification to Foster AI Literacy

Yasmine Belghith

Ph.D. Student in Human-centered Computing 

School of Interactive Computing 

Georgia Institute of Technology 

https://ybelghith.my.canva.site/

Date: Tuesday, January 20th, 2026 

Time: 12 PM - 2 PM Eastern Time (US)   

Location: TSRB 223 

Zoom: https://gatech.zoom.us/j/93000623000?pwd=YQ6hd5svLoOaL99kXcFW3AjunzJNaE.1

Committee 

Dr. Jessica Roberts (Advisor) - School of Interactive Computing, Georgia Institute of Technology

Dr. Betsy DiSalvo - School of Interactive Computing, Georgia Institute of Technology

Dr. Christopher MacClellan - School of Interactive Computing, Georgia Institute of Technology

Dr. Judith Uchidiuno - School of Interactive Computing, Georgia Institute of Technology

Dr. Duri Long - Department of Communication Studies, Northwestern University

Abstract 

Productive Failure (PF) is a two-phase pedagogical approach in which learners first engage in unguided problem solving before consolidating their learning by comparing their attempts to a canonical solution, a process shown to support conceptual understanding, transfer, and preparation for future learning. While PF has been well studied in instructor-led classroom contexts, its application to self-directed learning environments remains underexplored, particularly with respect to designing effective consolidation phases without expert facilitation. This proposal aims to investigate how PF can be operationalized within game-based learning environments, specifically digital educational escape rooms (DEERs), to support self-directed learning. Educational escape rooms (EERs) offer collaborative, puzzle-based structures that can align with PF principles, yet prior work suggests that their designs often lack strong theoretical grounding. This research aims to address this gap by leveraging PF as a guiding design framework.

Artificial intelligence (AI) literacy for non-technical young adults serves as the focal learning domain. Despite AI’s growing societal impact, many individuals who do not participate in computer science receive little formal instruction on AI concepts, leaving them vulnerable to misconceptions. PF-based EERs present a promising and accessible avenue for engaging diverse learners in AI literacy, supporting them in navigating AI’s abstract and novel concepts, and preparing them for future learning in this rapidly evolving technological domain. Through a mixed-methods program of studies, this work proposes to examine how a PF-based DEER can support AI literacy development among non-technical young adults. The expected contributions of this work include extending PF theory to self-directed learning contexts, developing analytic frameworks for examining PF processes without instructor facilitation, and empirically evaluating a PF-based DEER to inform the future design of theory-driven EERs.