On the Interplay of Human-AI Alignment,Fairness, and Performance Trade-offs in Medical Imaging

TL;DR

This paper systematically explores how human-AI alignment affects fairness and performance in medical imaging, showing that human insights can reduce biases and improve generalization, but require careful calibration to avoid trade-offs.

Contribution

It is the first to analyze the impact of human-AI alignment on fairness and robustness in medical imaging, highlighting the importance of balanced strategies.

Findings

Incorporating human insights reduces fairness gaps.

Human-AI alignment improves out-of-domain generalization.

Excessive alignment can cause performance trade-offs.

Abstract

Deep neural networks excel in medical imaging but remain prone to biases, leading to fairness gaps across demographic groups. We provide the first systematic exploration of Human-AI alignment and fairness in this domain. Our results show that incorporating human insights consistently reduces fairness gaps and enhances out-of-domain generalization, though excessive alignment can introduce performance trade-offs, emphasizing the need for calibrated strategies. These findings highlight Human-AI alignment as a promising approach for developing fair, robust, and generalizable medical AI systems, striking a balance between expert guidance and automated efficiency. Our code is available at https://github.com/Roypic/Aligner.