Want a robot life coach? Workers love it — as long as they don’t look like people

CAMBRIDGE, United Kingdom — Would you take life advice from a machine? According to a new study, robots are able to coach people on their mental health, provided these machines don’t look like humans.

University of Cambridge researchers deployed the well-being robot coaches in a tech consultancy firm – with one looking like a toy, and one looking like a “humanoid.” Every week, each bot coached 26 employees on how to feel good. Aside from their physical appearance, they were identical in terms of their voice, facial expressions, and coaching scripts.

Participants helped by the toy robot said they felt more of a connection than those who worked with the humanoid-like coach. Scientists put this down to popular culture leading us to believe automatons’ capacities are limited only by our imagination. In the real world, people were disappointed when their expectations of a human-like mental health machine were not met.

During the four-week experiment, the humanoid QTRobot (QT) or the toy-like Misty II robot (Misty) guided staff through four well-being exercises. The participants did not switch bots during the trial.

QT is a child-like humanoid roughly three feet tall, while Misty is a 36-centimeter-tall toy-like robot. Both have screens for faces that can be programmed with different facial expressions. In each session, the bot asked participants to recall a positive experience, or describe something they’re grateful for, before following up with questions.

After the session, employees assessed the robot’s performance in a questionnaire and interview. Participants reported the classes were useful and they would be open to talking to a robot in the future. Their main problems were expectations not being met, as they anticipated more interactivity.

woman sitting with robot mental health coach
Researchers from the University of Cambridge carried out a study in a tech consultancy firm using two different robot wellbeing coaches, where 26 employees participated in weekly robot-led wellbeing sessions for four weeks. Although the robots had identical voices, facial expressions, and scripts for the sessions, the robots’ physical appearance affected how participants interacted with it. CREDIT: University of Cambridge

“We interviewed different wellbeing coaches and then we programmed our robots to have a coach-like personality, with high openness and conscientiousness,” says study co-author Minja Axelsson in a media release. “The robots were programmed to have the same personality, the same facial expressions and the same voice, so the only difference between them was the physical robot form.”

“Our perceptions of how robots should look or behave might be holding back the uptake of robotics in areas where they can be useful,” Axelsson adds.

Despite the results, the scientists believe robots can be a useful tool to promote mental well-being in the workplace. The World Health Organization (WHO) recommends employers work to protect mental health in the office, but the practice is limited by a lack of resources and personnel. Robots show early promise that they could serve as the bridge. However, most well-being robot studies have been conducted in a laboratory, until now.

“We wanted to take the robots out of the lab and study how they might be useful in the real world,” says the paper’s first author Dr. Micol Spitale.

“It could be that since the Misty robot is more toy-like, it matched their expectations,” Spitale continues. “But since QT is more humanoid, they expected it to behave like a human, which may be why participants who worked with QT were slightly underwhelmed.”

Next, the team wants to develop the artificial coaches’ responsiveness during the sessions.

“The most common response we had from participants was that their expectations of the robot didn’t match with reality,” adds Professor Hatice Gunes from Cambridge’s Department of Computer Science and Technology, who led the research. “We programmed the robots with a script, but participants were hoping there would be more interactivity. It’s incredibly difficult to create a robot that’s capable of natural conversation. New developments in large language models could really be beneficial in this respect.”

The team presented the results at the ACM/IEEE International Conference on Human-Robot Interaction in Stockholm.

South West News Service writer Pol Allingham contributed to this report.

YouTube video