The life robotic

From left: Minja Axelsson, Nida Itrat Abbasi, Dr Micol Spitale, and Professor Hatice Gunes – researchers at Cambridge’s Affective Intelligence and Robotics Laboratory

From left: Minja Axelsson, Nida Itrat Abbasi, Dr Micol Spitale, and Professor Hatice Gunes – researchers at Cambridge’s Affective Intelligence and Robotics Laboratory

What is it about robots? Whether it’s the appeal of new technology, seeing science fiction become science fact, or exploring what they might be able to teach us about ourselves, robots and robotics seem to have an endless capacity to fire our imagination.

For Dr Micol Spitale, Nida Itrat Abbasi, and Minja Axelsson, researchers in Cambridge’s Affective Intelligence and Robotics Laboratory – directed by Professor Hatice Gunes - it’s the potential they have to help foster wellbeing in humans.

A recent study led by Dr Spitale, and co-authored by PhD student Minja, showed that robots could be useful as mental wellbeing coaches in the workplace, although perception of their effectiveness depends on what they look like. And separate research by PhD student Nida, and co-authored by Dr Spitale, suggested robots are potentially a promising tool in evaluating mental wellbeing issues in children, compared to parent-reported or self-reported testing.

It’s this fascinating faculty robots seem to have to elicit empathy in humans that drives the work of the team of roboticists. “Robots can work as virtual agents,” says Dr Spitale. “And for us it’s about collaborating with psychologists and wellbeing coaches to try and understand how we design these kinds of interactions to get the most out of the technology.”

Research at the Lab - in the University’s Department of Computer Science and Technology – suggests that children who might have wellbeing-related concerns, might view robots administering mental health tests as a friend or confidant, particularly if the robot is child-sized, and open up in a way that might not be possible with traditional methods.

“When children use screen-based tools, they often withdraw from the physical world,” says Nida. “But a robot is like a computer in physical form, and that’s one of the reasons they find them very engaging. Children are also very tactile, and they can touch the robots, and - in the case of ‘Nao’, the robot that helped assess them – even give it fist bumps!”

Dr Spitale – who began her postgraduate career in space engineering - moved into robotics, specifically the use of robots to support people with special needs, after seeing first-hand the connection they can make. Before starting her PhD she visited a lab researching robot therapy for children with autism, where non-verbal children were using an ‘expressive robot’, changing its eyes and mouth to communicate the emotions they themselves were feeling.

“The specific moment that changed everything for me was seeing how, after one of the children knocked the robot over, the other children stood it back up and gave it a sad face. I was like ‘Oh my gosh!’ This kind of tool can be used for children who don’t speak to express themselves. I fell in love with that idea; I could see the future of it - the impact it could have to help people.”

Nida – who also has a background in engineering – began her work to help vulnerable groups through her research when she was selected for a scholarship to study how cognitive training could support Alzheimer’s patients.

“That’s when I got interested in the idea of doing something to help people with technology. And so when I came to Cambridge for my PhD, I was fascinated by this project looking at the ability of robots to help with children’s wellbeing – especially the interdisciplinary aspect of it, and working with colleagues in the Department of Psychology.”

But if children can have an almost instinctive connection with robots, Dr Spitale’s research showed that for adults, a ‘wellbeing coach’ robot’s appearance could affect how they interact with it.

Participants in her study who did their wellbeing exercises with a toy-like robot said that they felt more of a connection with it than participants who worked with a humanoid-like robot - despite both robots being programmed with the same personality.

One of the reasons for this could be a mismatch between people’s idea of what a robot that looks humanoid should be capable of doing – based on popular culture, the media, and their experience of using virtual assistants like Alexa and Google Assistant – and the reality of the technology used in research labs, which doesn’t live up to their expectations.

“People’s perceptions of robots can depend on their age, their culture, their background,” says Dr Spitale. “And of course, it's very important to take these factors into account when you try to understand how the interaction could be successful between the human and the robot. With some of the adults in the study, we had to explain that the robot they were working with was actually ‘state-of-the-art’ in our field!”

Minja – who first became interested in robotics during her master’s degree, when she spent time at a consultancy exploring potential applications for an open source social robot, including assisting autistic children – says developing robots that meet people’s expectations was what designers call a ‘wicked problem’.

“People often imagine robots as being somewhere between computers and humans – and expect that they will incorporate elements of both in how they behave,” she says.

“They might expect a robot to be able to understand their body language, whereas they wouldn’t expect a computer to do that. At the same time, they might expect a robot to interact differently to a person, for example responding slower than a person would. So you’re trying to design something that fits neatly into all that – and that can be very interesting!”

But their potential to support wellbeing is there, says Minja, and that’s in part because of the ability robots have to elicit deep reflection in humans through the conversations they can facilitate.

“People taking part in these studies have said that just the fact that they're talking out loud is already doing a lot to help. And I think there's something to be said about the sense of presence a robot has. You’re engaged in social interaction, it might be with an object, but it still creates a different situation to what an app on a screen can do.

“One of the things I find most fascinating, in terms of our social interactions with robots, is how we suspend our disbelief. Even though we know it's a tool, we're able to still kind of jump into the social interaction in order for it to benefit us.”

Dr Spitale says parallel advancements in machine learning and AI will benefit the study of human-robot interaction, which began to develop as a field around 30 years ago.

“There is a resonance at the moment, and an appetite - in the media in particular - for these other spheres of computer science. For us it paves the way for more intelligent robots that can interact at a different level, and in the future, there may be opportunities for people to engage with effective, autonomous wellbeing coaches routinely in various spaces, for example in schools and workplaces. There could be an area where employees can work with the robots whenever they need, creating a visual reminder to complete a positive psychology exercise, or just to take time out.”

Nida added: “In schools and other childcare settings robots could be used as a tool to relieve pressure on existing resources, supporting the vital work of psychologists and clinicians, but, importantly, not replacing them or their expertise. We hope this technology can become available to anyone who can benefit from it.”

One of the unexpected outcomes of Nida’s recent research with children was the spontaneous creation of a physical message board by young participants who wanted to carry on their interaction with Nao after their assessment session had concluded.

“One of the children asked if they could write a message for Nao, and they used a marker on a whiteboard to say how they were feeling,” she says. “Then it just became a thing - and all the children were using the whiteboard to say ‘thank you’, and leave Nao their thoughts, and by the end, the whiteboard was filled.

“There is already research that shows children can consider robots as social companions and peers, so I’m so glad these children had the chance to engage with Nao. When I was a child, we didn’t have opportunities like this - to interact with a robot - but I would have loved it. We did have Meccano and Lego sets that my brother and I loved making things with. My father was an engineer, and I remember one of us made a robot figure with him using a kit and a Pepsi can. It still has pride of place, and I guess that was probably the start of all this for me!”

The text in this work is licensed under a Creative Commons Attribution 4.0 International License.

Researchers in the Affective Intelligence and Robotics Laboratory here are studying the potential robots have to help foster wellbeing in humans.