Human-Centred Control of AI Is Both Possible and Necessary
Joanna J. Bryson, NATO Science for Peace and Security Series: The Role of Transatlantic Cooperation in AI Human Oversight in Defence, Francisco Andrés Pérez (ed.) IOS Press, Amsterdam, and Springer Nature, Dordrecht, in press, expected 2026.
This was originally an invited talk for the meeting Clicking the pause: The role of Transatlantic cooperation in AI Supervision in May 2025. Both the talk and the book chapter were very much written to facilitate cooperation between US, EU, and other NATO members and partners, by helping them understand firstly that we can maintain human-centred control, and secondly, the relative strengths of both US and EU approaches to deploying and governing AI systems, particularly in military contexts. It contains some fairly direct language about working with NATO’s electoral autocracies, of which there are more than one.
Read More
A Moral Agency Framework for Legitimate Integration of AI in Bureaucracies
Chris Schmitz and Joanna J. Bryson, Sailing the Uncharted Waters: Public Administration and Emerging Technologies, Tan, Dan, and Sorin (eds.) Edwin Elgar Publishing, Cheltham, in press, expected 2026.
In Spring 2025 I realised that the ongoing myth of AI as necessarily opaque was actually related to the problem of transparency in government in interesting ways. Both governments and AI systems are subject to intentional design, so transparency can be a design criteria. Further, AI can be a method for achieving transparency in both, and democratic governance can ground that transparency in the requirements for legitimacy. Meanwhile, Chris Schmitz, a PhD student in public administration, recognised interesting correspondences between my theories of moral agency and a leading (Weberian) theory of bureaucracy. In this paper, we present a three-point Moral Agency Framework for legitimate integration of AI in bureaucratic structures: (a) maintain clear and just human lines of accountability, (b) ensure humans whose work is augmented by AI systems can verify the systems are functioning correctly, and (c) introduce AI only where it doesn’t inhibit the capacity of bureaucracies towards either of their twin aims of legitimacy and stewardship. We suggest that AI introduced within this framework can not only improve efficiency and productivity while avoiding ethics sinks, but also improve the transparency and even the legitimacy of a bureaucracy. Earlier version appeared in Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, 8(3), 2292-2293, Spring 2025. The 2026 version has more direct comments about governments that shift from being electoral democracies to being electoral autocracies.
Read More
