July 21, 2025

On Wednesday, July 16, 2025, Theresa Willem successfully defended her dissertation "Embedding ethics in diagnostic imaging AI. Navigating social and ethical risks by analyzing their origins in technical features and social processes".

In recent decades, medical AI has experienced an enormous boom. At the same time, medical ethical and social problems have emerged that can be attributed to AI. However, as a scientific community, we struggle to anticipate these problems and act before they disadvantage individuals or groups. Looking at AI in diagnostic imaging, Dr. des. Willem asked: "How can the ethical and social risks of emerging diagnostic imaging AI be effectively anticipated and embedded in ongoing research and development?" . Her research combines biomedical ethics, AI/technology ethics, science and technology studies, critical data studies and computer science.

This work was the first at the Institute for the History and Ethics of Medicine to implement the jointly developed Embedded Ethics approach and is thus exemplary for the deep interdisciplinary cooperation between the different schools.

The term "embedded ethics" describes an approach in which the development of new technologies (especially in the field of disease and health, with a focus on AI and machine learning) is accompanied from the outset by a deep integration of ethical and social methodology.

Dr. des. Willem was deeply involved in the development processes in dermatology and radiology at the TUM University Hospital and at the Helmholtz Institute and has not only made a significant contribution to increasing the quality of products and results at all these institutions, but above all to ensuring that the development process took ethical and social implications into account and addressed them directly in a responsible manner.

In addition, important basic theoretical concepts of the embedded ethics approach have been further developed in her dissertation.

The work has led to a large number of high-ranking publications, including:

• Willem, T., Krammer, S., Böhm, A. S., French, L. E., Hartmann, D., Lasser, T., & Buyx, A. (2022). Risks and benefits of dermatological machine learning health care applications-An overview and ethical analysis. Journal of the European Academy of Dermatology and VeneReology, 36(9), 1660-1668.
• Willem, T., Wollek, A., Cheslerean-Boghiu, T., Kenney, M., & Buyx, A. (2025). The social construction of categorical data: Mixed methods approach to assessing data features in publicly available datasets. JMIR Medical Informatics, 13, e59452.
• Willem, T., Shitov, V. A., Luecken, M. D., Kilbertus, N., Bauer, S., Piraud, M., ... & Theis, F. J. (2025). Biases in machine-learning models of human single-cell data. Nature Cell Biology, 1-9.
• Buyx, A., McLennan, S., Müller, R., Tigard, D., Fiske, A., Meier, L. J., ... & Willem, T. (2024). Embedded Ethics in Practice: A Toolbox for Integrating the Analysis of Ethical and Social Issues into Healthcare AI Research: Embedded Ethics in Practice: A Toolbox.