This research analyzes potential demographic biases in foundation models used for X-ray image classification, with critical implications for equitable healthcare delivery.

• CLIP-based models trained on medical imaging show promise for diagnostic accuracy but may perpetuate hidden biases
• Research evaluates fairness across demographic groups (age, sex, race) in medical image classification
• Findings reveal important performance disparities that must be addressed before clinical deployment
• Study provides a framework for systematic bias evaluation in medical AI systems

Why it matters: As AI increasingly supports medical decision-making, ensuring these systems perform equitably across all patient populations is essential for responsible healthcare innovation and patient safety.

Fairness Analysis of CLIP-Based Foundation Models for X-Ray Image Classification