Research

Human control and accountability in autonomous weapons

My main research asks what happens to human judgement and responsibility when weapon systems act faster than the people meant to supervise them. It began with my MSc dissertation at the University of Glasgow, a case study of unintended engagements by the U.S. Patriot air-defence system in the Iraq War.

Two current papers develop this argument. Command Under Automation examines the limits of human control in automated air defence. The Supervisory Illusion asks whether meaningful human control, as it is usually defined, gives operators real authority or only the appearance of it, and what that means for responsibility when automated systems fail.

I am extending this work in two directions. One project uses open-source records of drone components to test how far component provenance can support the attribution of strikes by autonomous drones. The other compares the defence procurement documents of the UK, the EU, the US and India with the positions these states take in UN discussions on autonomous weapons, asking whether the documents actually specify human control.

Counter-terrorism law and preventive detention in India

My second strand grows out of legal research in Indian chambers. Since 2019, my professional legal research has supported matters involving preventive detention, counter-terrorism and financial crime in India, and that work shapes the questions I ask as a scholar. I study how counter-terror law operates before any trial begins: how preventive detention and restrictive bail standards turn pre-trial custody into a form of punishment, and how this narrows the space for civic and political life. I develop this argument in Punishment Before Conviction, a manuscript currently under external peer review.

I am also developing a comparative project on secret evidence and judicial deference, setting Indian sealed-cover practice alongside closed material procedure in the United Kingdom.

AI, governance and geopolitics

My third strand asks how algorithms change the exercise of power, both within the state and between states. In a co-authored article on India, we propose a Substantive Discretion Test for judging when automated systems in welfare and tax administration have taken over decisions that the law entrusts to officials. A related manuscript examines how AI-driven resource allocation can exclude communities in climate-vulnerable states.

Between states, I study how Indo-Pacific middle powers manage their dependence on American and Chinese AI without either full alignment or self-sufficiency. I also work on generative AI and influence operations. Drawing on the influence networks that OpenAI has disclosed since 2024, I argue that generative AI has made content cheap without making audiences easier to reach, so access to distribution remains the real barrier.