Research Overview

I am a PhD researcher in Design at Lancaster University, in the Faculty of Arts and Social Sciences, supervised by Dr Louise Mullagh and Dr Elisavet Christou. My work covers AI governance, public-facing AI systems, and the epistemic infrastructures of public knowledge institutions: museums, libraries, archives, universities and public broadcasters. I am interested not only in how these institutions deploy generative AI, but in how they disclose it. Upstream infrastructures such as catalogue records, data visibility and disclosure practices quietly shape what an AI system can retrieve, rank and re-describe, and I treat disclosure itself as a piece of governance infrastructure: once an institution tells the public that an AI was involved, I ask how it routes accountability, and how far its technical workings and sources can actually be inspected. My method is deliberately external. I run researcher-led public audits and inspect disclosure practices, building arguments from governance traces anyone can observe rather than from institutional interviews or insider access.

My doctoral project examines the AI-mediated interpretation of contested cultural heritage, with Chinese collections in European museums, including Yuanmingyuan (Old Summer Palace) objects, as its case. I trace how an Orientalist framing can be written into a catalogue description, carried forward through platform standardisation, and end up as the raw material an AI system later retrieves and re-describes when it generates an interpretation, for instance in retrieval-augmented systems. The study works across three connected layers: an audit of the catalogue substrate that AI draws on, controlled testing of how AI outputs shift under different source-governance conditions, and the design of practical tools, including a disclosure matrix and a catalogue-to-AI risk checklist. Around this case I also work on cross-model audits of how generative systems route authority and source boundaries, on responsible-AI signalling and "responsibility-washing" in commercial products, and on the UK's position between US-style AI safety testing and European AI regulation.