AI in Medicine: Robotics & Surgery
Explainability and Reliability of AI for Medical Imaging
Prof. Hristina Uzunova
Artificial intelligence (AI) is playing an increasingly important role in robot-assisted interventions, particularly in medical imaging for precise surgical planning, intraoperative navigation, and postoperative assessment. Explainability and uncertainty analysis are often cited as key enablers of trustworthy AI. Explainable AI promises to make algorithmic decisions more transparent, yet many current methods risk producing oversimplified or even misleading insights that may not translate into genuine clinical understanding. Similarly, uncertainty analysis offers valuable information about prediction reliability, but methods remain computationally demanding and their interpretation is not always intuitive for end-users. This talk explores both the opportunities and the limitations of explainability and uncertainty analysis for AI in medical imaging.