Abstract

Artificial Intelligence (AI) is widely promoted as a transformative force in healthcare, with the potential to revolutionize diagnostics, treatment, and administration. Yet in practice, scaling efforts struggle as projects move from pilot initiatives to broader clinical integration. As a result, we lack insight into why the same AI system may be embraced in one hospital but resisted or abandoned in another. To address this knowledge gap, this study explores how and why the meaning of an AI tool can shift during scaling, drawing on an illustrative case study of automated AI-enabled assessment for a disease in a European country. We propose a conceptual framework to examine how stakeholder interpretations change over time and how these shifts affect the success of AI-scaling. The findings show that AI-scaling is not a simple matter of scaling a working system, but an ongoing process of negotiation about responsibility and risk. Theoretically, the paper contributes to IS research by highlighting the value of temporally aware, interpretive approaches to understanding AI-scaling. Practically, it offers six lessons learned and six steps for more successful AIscaling. These can be useful for healthcare implementers seeking to maintain alignment across diverse stakeholders and evolving institutional landscapes.

Download

Share

COinS