Areas of interest

My work draws on behavioural science to understand how AI changes the conditions under which people think, decide, and take responsibility. This includes how people adapt to AI in use, how systems are designed around that behaviour, how AI systems behave over time in interaction, and how organisations translate these dynamics into oversight and accountability.

As AI systems become part of everyday work, the questions are no longer only about what the systems can do. They are about how people rely on them, how that reliance changes behaviour over time, how systems are designed around those behaviours, and how organisations remain accountable when decisions are distributed across humans and machines.

My background in behavioural science is useful here because it provides a way of making these changes visible. It offers a vocabulary for describing how people adapt, where judgement breaks down, and why apparently simple ideas such as trust or oversight often fail in practice.

Across these areas, the common thread is the same: AI does not only change what work is done. It changes how judgement is exercised, how decisions are structured, and how responsibility is distributed. Understanding those changes requires looking at people, systems, and organisations together.

I am interested in what AI does to human judgement over time. This includes reliance and over-reliance, shifts in attention and effort, cognitive offloading, and how repeated interaction changes confidence, motivation, and sense of agency.

Behavioural science provides the language to describe these changes in a precise way: how trust forms and breaks, how judgement is supported or weakened, and how patterns of use accumulate into something more than a series of isolated decisions.

These behavioural patterns shape what good design requires. I am interested in how behavioural knowledge translates into design decisions: how systems support or undermine judgement, how feedback is structured, where friction is useful, and how decision support aligns with how people actually think rather than how we assume they do.

The aim is to avoid a common failure mode where systems work in principle but create problems in use because they are not built for real human behaviour.

AI systems are often treated as tools, but in many contexts they exhibit patterns that are only visible through interaction. I am interested in how these patterns can be studied: how systems behave across contexts, how they adapt, and what emerges over repeated use or in multi-agent settings.

This perspective treats AI not only as something to build or regulate, but as something to observe and understand empirically.

These questions become more consequential when AI affects decisions with real outcomes. I am interested in how organisations translate behavioural realities into governance: how responsibility is defined, how oversight is expected to work, and whether those expectations are realistic given how people actually behave.

This includes accountability, escalation, distribution of risk, and whether arrangements around AI can withstand scrutiny when something goes wrong. Behavioural science helps ground these questions in actual behaviour rather than abstract principles.