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

Haozhe Luo, Ziyu Zhou, Zixin Shu, Aurélie Pahud de Mortanges, Robert Berke +1 more
5/15/2025
cs.CVcs.AIcs.CL

Official code: Roypic/Aligner (0 stars, last commit 2025-09-23).

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.

View on arXivView PDF

Code Implementations(1)

Roypic/AlignerOfficial100%
00Python, Jupyter NotebookMay 15, 20251 years ago

Similar Papers

Towards objective and systematic evaluation of bias in artificial intelligence for medical imaging

87% match1 repo122023

FairDomain: Achieving Fairness in Cross-Domain Medical Image Segmentation and Classification

85% match1 repo392026

Evaluating and Mitigating Bias in AI-Based Medical Text Generation

85% match1 repo242025

Unveiling Fairness Biases in Deep Learning-Based Brain MRI Reconstruction

85% match1 repo32023

Data-Algorithm-Architecture Co-Optimization for Fair Neural Networks on Skin Lesion Dataset

85% match2024

Evaluating the Fairness of Neural Collapse in Medical Image Classification

84% match2024

Cite this paper

@article{luo2025interplay,
  title  = {On the Interplay of Human-AI Alignment,Fairness, and Performance Trade-offs in Medical Imaging},
  author = {Haozhe Luo and Ziyu Zhou and Zixin Shu and Aurélie Pahud de Mortanges and Robert Berke and Mauricio Reyes},
  year   = {2025},
  eprint = {2505.10231},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2505.10231v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2505.10231.svg)](https://codeofpaper.com/paper/2505.10231)

Discussion