A Learning Approach for Increasing AI Literacy via XAI in Informal Settings

To achieve AI literacy, the AI community employs explainable AI (XAI), to increase AI literacy for those outside of formal educational settings. Designing and evaluating XAI remains an open question that can be guided by existing learning science research. When designers view their XAI through a learning lens, they may better define, assess, and compare explanation implementations. We surveyed and interviewed designers of interactive explanations for AI to identify how practitioners build their XAI and to better understand how a learning lens can be applied for explanations of complex AI concepts.