Looking into Pandora’s Black Box: A Systematic Literature Review to Understand Transparency in Human-Agent Interaction

Transparency has emerged as a critical construct in human-agent interaction (HAI) research. Given the interdisciplinary field of HAI, researchers from different backgrounds tend to study transparency in different ways, yet there remains no clear consensus on how agent transparency should be defined, measured, or manipulated. Therefore, this systematic literature review takes a crucial step toward addressing this challenge by synthesizing how transparency has been conceptualized and operationalized across HAI research. Through this comprehensive synthesis, we examined how transparency has been defined, measured, and manipulated across studies. Our findings reveal that transparency is most commonly defined as information disclosure, yet definitions, measures, and manipulations of transparency remain highly inconsistent across studies. Notably, most papers that manipulated transparency never verified whether participants actually perceived the system as more transparent, highlighting a critical gap between designed and perceived transparency and the need for a validated measure of perceived transparency.