Sciencepulse

Helen Hastie Wants Robots You Can Talk To, and Trust

Robotics·October 26, 2026

Helen Hastie Wants Robots You Can Talk To, and Trust

What if robots were not lone machines carrying out isolated jobs, but teammates you could actually hold a conversation with? That is the question driving Helen Hastie, Professor of Human-Robot Interaction and Head of the School of Informatics at the University of Edinburgh. In a recent episode of the BBC Radio 4 series The Life Scientific, presenter Jim Al-Khalili spoke with her about what it would take for machines to respond naturally to conversational cues and explain their own actions.

Hastie's route into the field runs through language technology. She began by building early dialogue systems, the distant ancestors of today's generative AI chatbots. Her later work has moved those ideas into physical settings. Her team has developed bots that can serve coffee and make small talk, systems that help struggling pupils with a patient, empathetic tone, and prototypes that support triage nurses by offering calm, confident guidance. She has also taken a leading role in several of the UK's flagship robotics programmes, including as co-lead of the National Robotarium.

The conversation includes a memorable detour. Al-Khalili admits that an early chatbot once scolded him for being rude, a small moment that shows how quickly people bring social expectations to software that talks back. Hastie uses that instinct as a starting point. If people treat conversational machines as social partners, designers have to take manners, timing and clarity seriously.

For Hastie, the central goal is trust. A robot that is useful but unpredictable will not last long in a classroom, a care setting or a hospital ward. She wants machines that are reliable, that can say why they chose a particular action, and that fit naturally into the places where people work and live. The aim, as she puts it in the episode, is not simply robots that exist in our world, but ones we are glad to have around.

That framing marks a shift in how robotics is discussed. For years the field was measured by what machines could physically do. Hastie's work suggests the harder problem is social: whether people will understand, rely on and accept the systems they share space with. The interview is a useful reminder that the next stage of robotics may be judged less by dexterity than by whether a machine can be a credible colleague.

Reporting based on an external source.