Íñigo de Troya is a postdoctoral researcher at the Oxford Internet Institute, working with Prof. Luc Rocher in the Synthetic Society Lab. Íñigo’s research examines public sector AI systems and their governance through a sociotechnical systems lens. Drawing on safety science, critical AI studies, and participatory design, he studies how power relations shape how AI safety is negotiated and realised across the technical, operational, organisational, and institutional contexts in which these systems are embedded. Íñigo is particularly interested in how civil society actors can inform design and governance practices that serve the public interest, and how regulatory bodies can learn from the methods and history of safety science.
His research has examined risk prediction systems in public employment services; participatory methods for mapping sociotechnical AI systems; municipal algorithm registers; value alignment of Large Language Models; the role of civil society organisations in AI governance; the perspectives of civil servants using decision-support tools in public services; and more.
Íñigo completed his PhD at the Sociotechnical AI Systems Lab in TU Delft’s Faculty of Technology, Policy, and Management, under the supervision of Profs. Roel Dobbe, Neelke Doorn, and Virginia Dignum. Prior to TU Delft, Íñigo was a data scientist, research fellow, and infrastructure engineer at the Data Science for Social Good Foundation, based at Nova SBE in Lisbon, Portugal. Before turning to public policy, he worked on machine listening applications for home security and birdwatching. He received a MEng in Electrical & Electronic Engineering from Imperial College London, and a MRes in Neuroinformatics & Computational Neuroscience at The University of Edinburgh.
AI governance; public sector AI; responsible AI; power; sociotechnical systems; system safety; safety science; AI safety
The Synthetic Society Lab aims to improve how we research, access, and use sensitive human data. The fellowship will advance an evidence-based science of privacy engineering to make research using digital traces safe and reliable.
