Validated AI systems and their struggle to scale in healthcare

Join our Operations and Technology Management seminar

Speaker: Professor Uwe Fischer, Yale School of Medicine

About the seminar topic

The talk will explore why validated AI systems often struggle to scale in routine clinical practice despite strong technical performance and growing regulatory approval.

The Operations and Technology Management seminar in collaboration with the Cambridge Centre for AI in Medicine will examine the hypothesis that AI may differ from previous healthcare technologies because it increasingly contributes clinical judgment rather than simply providing information. Whether AI is viewed as fundamentally different or as the next stage in healthcare technology, the practical question remains the same: are healthcare organisations prepared to absorb, govern and operationalise AI-generated recommendations at scale?

We will examine the growing gap between AI development and routine clinical adoption, introduce the concept of institutional absorption capacity and consider how workflow integration, governance and liability influence implementation success. Finally, we will discuss the potential role of simulation-based approaches to evaluate implementation readiness before deployment.

The objective is to move the discussion beyond model performance and toward the organisational conditions required for sustainable AI adoption in healthcare.

Speaker bio

Uwe Fischer, MD/PhD, is a vascular surgeon and Assistant Professor of Surgery at Yale School of Medicine. His work focuses on the intersection of AI, clinical decision-making, healthcare implementation and vascular surgery.

His recent research examines why validated AI systems frequently fail to achieve widespread adoption despite strong technical performance. He has proposed the concept of institutional absorption capacity, a framework for understanding how healthcare organisations integrate, govern and operationalise AI-generated clinical decisions. His current work explores the use of simulation-based approaches to prospectively assess implementation readiness before deployment.

Register

No registration required.