A year ago, if you'd asked me what my PhD would be about, I would have given you a tidy, one-sentence answer. Now that I'm actually in it, the honest answer is messier — and better. I'm researching Human-AI Interaction, but the question underneath all of it is simpler than it sounds: can we build AI systems that actually reflect human values, rather than just performing them?
Why this question, and why now
Before starting my PhD at the University of Manchester, I spent time on both sides of this problem — analysing data for AI systems that were meant to help people, and studying, from a psychology background, how people actually perceive and judge the actions of others. Somewhere between those two experiences, a pattern kept showing up: the systems that failed weren't failing on accuracy. They were failing on trust. People didn't reject AI because it was wrong; they rejected it because they couldn't tell why it was doing what it was doing, or whether it had their interests in mind at all.
That gap — between what a system can technically do and what a person can actually understand, question, and trust — is where my research lives. My supervisors, Prof. Caroline Jay and Dr. Emily C. Collins, have been pushing me to treat that gap not as a UX problem to smooth over, but as an ethical one to take seriously.
What I'm actually working on
Right now, that means looking closely at moral agents, software engineering workflows, and the everyday decisions that shape whether an AI system ends up being honest with the people who rely on it. It also means going back to first principles: what does it even mean for an explanation to be honest? Whose values get encoded when we decide what "ethical AI" looks like in practice? These aren't questions with clean answers, which is exactly why they're worth a PhD's worth of attention.
What's next
I'll be writing here as the research develops — the ideas that hold up, the ones that don't, and the occasional conference talk that changes how I think about all of it. If any of this resonates with your own work, or you just want to argue with me about it, my inbox is open.