Robot wellbeing coaches work... if they look like a toy

ROBOTS can improve mental wellbeing in the workplace, but they have to look right, a Cambridge University report suggests.

A study in a tech consultancy firm using two different robot wellbeing coaches found people felt more of a connection with toy-like robots than humanoid ones.

Perception of robots is affected by popular culture, where the only limit on what robots can do is the imagination, researchers said.

When people are faced with a robot in the real world, however, it often does not live up to expectations.

According to the researchers, since the toy-like robot looks simpler, people may have had lower expectations and ended up finding the robot easier to talk to and connect with.

Those who worked with the humanoid robot found their expectations did not match reality, since the robot was not capable of having interactive conversations, they said.

The researchers collaborated with a local technology company, Cambridge Consultants, to design and implement a workplace wellbeing programme using robots. Over the course of four weeks, 26 employees were guided through four different wellbeing exercises by one of two robots: either the Qtrobot (QT) or the Misty II robot (Misty). The QT is a childlike humanoid robot and roughly 90cm tall, while Misty is a 36cm tall toylike robot. Both have screen faces that can be programmed with different facial expressions.

Dr Micol Spitale, the paper’s first author, said: “It could be that since the Misty robot is more toy-like, it matched their expectations. But since QT is more humanoid, they expected it to behave like a human, which may be why participants

‘It’s incredibly difficult to create a robot that’s capable of natural conversation’

who worked with QT were slightly underwhelmed.”

After speaking to wellbeing coaches, the researchers programmed the robots to have a coach-like personality, with high openness.prof Hatice Gunes, from Cambridge’s Department of Computer Science and Technology, who led the research, said: “It’s incredibly difficult to create a robot capable of natural conversation. New developments in large language models could be beneficial.”

The findings are presented at the ACM/IEEE International Conference on Human-robot Interaction in Stockholm.