# Are generative models fair? A study of racial bias in dermatological image generation

arXiv: https://arxiv.org/abs/2501.11752 · HTML: https://mlsanity.com/paper/2501.11752/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 25 references that are in the map.
- First posted on arXiv 20 Jan 2025.

## Builds on (most cited first, 14 of 14)

- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Hidden Stratification Causes Clinically Meaningful Failures in Machine Learning for Medical Imaging](https://mlsanity.com/paper/1909.12475/) (2019)
- [Trustworthy AI: From Principles to Practices](https://mlsanity.com/paper/2110.01167/) (2021)
- [Evaluating Deep Neural Networks Trained on Clinical Images in Dermatology with the Fitzpatrick 17k Dataset](https://mlsanity.com/paper/2104.09957/) (2021)
- [Reading Race: AI Recognises Patient's Racial Identity In Medical Images](https://mlsanity.com/paper/2107.10356/) (2021)
- [Disparities in Dermatology AI: Assessments Using Diverse Clinical Images](https://mlsanity.com/paper/2111.08006/) (2021)
- [IntroVAE: Introspective Variational Autoencoders for Photographic Image Synthesis](https://mlsanity.com/paper/1807.06358/) (2018)
- [RadFusion: Benchmarking Performance and Fairness for Multimodal Pulmonary Embolism Detection from CT and EHR](https://mlsanity.com/paper/2111.11665/) (2021)
- [Improving dermatology classifiers across populations using images generated by large diffusion models](https://mlsanity.com/paper/2211.13352/) (2022)
- [Consensus and Subjectivity of Skin Tone Annotation for ML Fairness](https://mlsanity.com/paper/2305.09073/) (2023)
- [Feature robustness and sex differences in medical imaging: a case study in MRI-based Alzheimer's disease detection](https://mlsanity.com/paper/2204.01737/) (2022)
- [Slicing Through Bias: Explaining Performance Gaps in Medical Image Analysis using Slice Discovery Methods](https://mlsanity.com/paper/2406.12142/) (2024)
- [Revisiting Skin Tone Fairness in Dermatological Lesion Classification](https://mlsanity.com/paper/2308.09640/) (2023)
- [PASSION for Dermatology: Bridging the Diversity Gap with Pigmented Skin Images from Sub-Saharan Africa](https://mlsanity.com/paper/2411.04584/) (2024)

## Built on by (most cited first, 7 of 7)

- [Neutralizing the Narrative: AI-Powered Debiasing of Online News Articles](https://mlsanity.com/paper/2504.03520/) (2025)
- [DermaCon-IN: A Multi-concept Annotated Dermatological Image Dataset of Indian Skin Disorders for Clinical AI Research](https://mlsanity.com/paper/2506.06099/) (2025)
- [The Multimodal Paradox: How Added and Missing Modalities Shape Bias and Performance in Multimodal AI](https://mlsanity.com/paper/2505.03020/) (2025)
- [Coffee: Controllable Diffusion Fine-tuning](https://mlsanity.com/paper/2511.14113/) (2025)
- [Bayesian generative models can flag performance loss, bias, and out-of-distribution image content](https://mlsanity.com/paper/2503.17477/) (2025)
- [Equity Bias: An Ethical Framework for AI Design](https://mlsanity.com/paper/2604.21907/) (2026)
- [A Data-Interventional Framework for Auditing Privacy and Fairness in Generative Medical Imaging](https://mlsanity.com/paper/2609.26623/) (2026)

## Abstract (arXiv)

Racial bias in medicine, such as in dermatology, presents significant ethical and clinical challenges. This is likely to happen because there is a significant underrepresentation of darker skin tones in training datasets for machine learning models. While efforts to address bias in dermatology have focused on improving dataset diversity and mitigating disparities in discriminative models, the impact of racial bias on generative models remains underexplored. Generative models, such as Variational Autoencoders (VAEs), are increasingly used in healthcare applications, yet their fairness across diverse skin tones is currently not well understood. In this study, we evaluate the fairness of generative models in clinical dermatology with respect to racial bias. For this purpose, we first train a VAE with a perceptual loss to generate and reconstruct high-quality skin images across different skin tones. We utilize the Fitzpatrick17k dataset to examine how racial bias influences the representation and performance of these models. Our findings indicate that VAE performance is, as expected, influenced by representation, i.e. increased skin tone representation comes with increased performance on the given skin tone. However, we also observe, even independently of representation, that the VAE performs better for lighter skin tones. Additionally, the uncertainty estimates produced by the VAE are ineffective in assessing the model's fairness. These results highlight the need for more representative dermatological datasets, but also a need for better understanding the sources of bias in such model, as well as improved uncertainty quantification mechanisms to detect and address racial bias in generative models for trustworthy healthcare technologies.