A hospital deploys an AI system that recommends which patients should be prioritized for a specialist consultation. Clinicians complain that they cannot understand why the model flags certain patients, making them reluctant to trust its recommendations. Which responsible AI principle should the development team focus on to address this concern?