Abstract: Concerns regarding the power dynamics underlying AI development and deployment have increasingly gained attention. However, it is unclear how to conceptually understand these relations which inhibits their normative assessment. In this dissertation, I analyze these power dynamics using the neo-republican conception of freedom as non-domination. Domination occurs when someone is subject to an arbitrary and uncontrolled power. Such relations of domination occur in the context of AI systems applied in core societal sectors between the ‘shapers’ of AI systems (developers and deployers) and those ‘affected’ (the people subject to the AI system). The domination is grounded in at least three reasons: (1) the distribution of decision-making power, (2) technical limitations of AI systems that result in accountability gaps, and (3) underlying societal structures that enhance the power of the shapers and inhibit the power of the stakeholders. To safeguard freedom in the digital age, I argue that AI development requires the explicit intention to mitigate relations of domination. In other words, we must design for non-domination. This necessarily requires considering the broader societal context, such as current regulatory initiatives and the political economy.

Publications:

• Maas, J. (2022). A Neo-Republican Critique of AI Ethics. Journal of Responsible Technology, 9, 100022. https://doi.org/10.1016/j.jrt.2021.100022(Part of this dissertation)
• Veluwenkamp, H., Capasso, M., Maas, J., & Marin, L. (2022). Technology as driver for morally motivated conceptual engineering. Philosophy & Technology, 35(3), 71. https://doi.org/10.1007/s13347-022-00565-9(Part of this dissertation)
• Maas, J. (2023). Machine learning and power relations. AI & SOCIETY, 38(4), 1493-1500. https://doi.org/10.1007/s00146-022-01400-7 (Part of this dissertation)
• Stockinger, E., Maas, J., Talvitie, C. & Dignum, V. (2024). Trustworthiness of Voting Advice Applications in Europe. Ethics and Information Technology.
• Maas, J., & Inglés, A. M. (2024). Beyond Participatory AI. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (Vol. 7, pp. 932-942). (Part of this dissertation)
• Maas, J. & Van den Hoven, J. (forthcoming). Opening the Black Box of AI, Only to be Disappointed. In Computer Ethics Across Disciplines – Applying Deborah Johnson’s Philosophy to Algorithmic Accountability and AI. Noorman, M. & Verdicchio, M. (Eds.). Springer. (Part of this dissertation)