Abstract
Design approaches such as Value Sensitive Design (VSD) facilitate the alignment of technology design with societal and ethical values. VSD serves as a proactive and iterative framework for integrating ethical considerations into design, encompassing three interrelated types of investigations: conceptual investigations, which analyze relevant values and stakeholders; empirical investigations, which gather data about stakeholders’ experiences and value prioritization; and technical investigations, which focus on embedding these values into the functionality of technologies. This paper critically examines the application of VSD for artificial intelligence (AI) through a systematic literature review. By addressing VSD’s theoretical pillars and AI’s ambiguous boundaries, the study maps these complexities to the descriptions of the practical implementation in value sensitive AI (VSAI) literature. Our findings contribute to a comprehensive framework identifying practical and theoretical gaps, such as the operationalization of VSD’s tripartite methodology in AI contexts and value prioritization. We contribute a synthesis of recommendations from the literature as well as our own to fill these gaps and offer actionable strategies for future VSAI applications. Together, these insights clarify the current methodological landscape, and provide researchers with practical guidance in building value sensitive AI.