[1] Y. Elias, T. P. Humbert, L. Olson and E. Guzmán, "What is Unethical About Software? User Perceptions in the Netherlands," 2025 IEEE/ACM 47th International Conference on Software Engineering: Software Engineering in Society (ICSE-SEIS), Ottawa, ON, Canada, 2025, pp. 118-129, doi: 10.1109/ICSE-SEIS66351.2025.00017.

Abstract: Software has the potential to improve lives. Yet, unethical and uninformed software practices are at the root of an increasing number of ethical concerns. Despite its pervasiveness, few research has analyzed end-users perspectives on the ethical issues of the software they use. We address this gap, and investigate end-user's ethical concerns in software through 19 semi-structured interviews with residents of the Netherlands. We ask a diverse group of users about their ethical concerns when using everyday software applications. We investigate the underlying reasons for their concerns and what solutions they propose to eliminate them. We find that our participants actively worry about privacy, transparency, manipulation, safety and inappropriate content; with privacy and manipulation often being at the center of their worries. Our participants demand software solutions to improve information clarity in applications and provide more control over the user experience. They further expect larger systematic changes within software practices and government regulation.

[2] Ö. Karaçam, T. P. Humbert and E. Guzmán, "Uncovering Patterns in Users' Ethical Concerns About Software," 2024 IEEE 32nd International Requirements Engineering Conference (RE), Reykjavik, Iceland, 2024, pp. 466-474, doi: 10.1109/RE59067.2024.00055.

Abstract: Ethical concerns about software applications, e.g., worries about privacy breaches, user manipulation, and discrimination, have gained prominence recently. Research shows that users voice these concerns in app reviews and that they can be detected using machine learning and deep learning techniques. These techniques usually operate as black-boxes, making it difficult to understand the context of users' ethical concerns. We address this issue by presenting a transparent approach that uses pattern mining and graph theory to yield additional context to the ethical concern classifications made by machine learning algorithms. We compare a simple frequent pattern mining and a high-utility mining algorithm and assess the resulting rules through commonly used metrics. Finally, we visualize and interpret preliminary results in an interactive graph. We mined 3,101 reviews of ten popular apps mentioning diverse ethical concerns and present the results for two apps in detail. Our results show that pattern mining algorithms and graph visualizations are promising directions for detecting contextual information of ethical concerns about software. This work is a step toward ensuring that ethical concerns are methodically thought through and integrated into the software development life cycle.