A hospital is building an AI system that detects skin conditions from patient photographs. During testing, the team discovers the model performs well for lighter skin tones but poorly for darker skin tones because the training images mostly featured lighter-skinned patients. Which responsible AI principle is the team failing to uphold?