Our research.

We study how AI systems use context, interact with people, and recognise when to ask for help.

Questions we’re exploring.

Context & memory: When should an assistant remember a detail, ask again, or forget it? We explore how people can inspect and correct the context a system carries into its next conversation.
Human interaction: How does someone tell that a digital face is paying attention? Small changes in gaze, timing, and expression give us specific cues to study.
Trustworthy intelligence: Where can a tool do a task more reliably than a person, and where does it need oversight? We compare errors on the same task and examine when a system should hand a decision back to its user.

Our website prototype uses one face to explore how a glance, a blink, and the beginnings of a smile affect the impression of a face.