How Robots Can Help Kids Develop Social Skills

The following is an excerpt from The New Breed: What Our History With Animals Reveals About Our Future With Robots by Kate Darling.

Buy The Book

The New Breed: What Our History With Animals Reveals About Our Future With Robots

Buy

The Huggable is a soft, fully animated blue teddy bear with a green nose that wrinkles in mischief and delight. The robot was developed by the MIT Media Lab’s Personal Robots Group. The researchers collaborated with Boston Children’s Hospital to bring the Huggable to pediatric patients and their families with the goal of mitigating the children’s stress, anxiety, and pain by engaging them in playful interactions. While pediatric doctors and nurses are fairly skilled in talking to children, using the (in this case remotely controlled) animated Huggable bear in tandem with a care provider contributed to a better social and emotional experience for the patients. That the presence of a robot can make children more comfortable talking to the doctors and nurses is no surprise. One of the benefits we’ve seen in animal-assisted therapy has been in giving patients a type of social support that humans can’t provide.

Robots are unlike some of the other interventions we’ve used with kids previously. Static toys and screens have different characteristics. Robots like the Huggable can talk, move their bodies, and have animated facial expressions. Unlike a passive stuffed animal or a flat animation on an iPad, these robots interact in the physical world.

Human-robot interaction (HRI) research shows that this lets the robots communicate with children on a more active and socially present level. HRI researchers are also studying how we might engage children in learning with robot teaching assistants, and with much better results than computers. The robots can adjust their responses to the child’s learning needs, personalizing the programs to the parts the child is struggling with, all the while using the physical interaction that works so much better than a screen. A child may enjoy interacting with the robot at first because it’s fun and novel for them. But the robots are showing promise at getting kids to work hard and overcome learning difficulties beyond their initial excitement and fascination with the new device.

Like animals, these robots can’t replace human teachers. In a world of overcrowded classrooms, where one-on-one interactions with teachers are too rare, it’s possible that they could provide better than nothing. But their ideal and most promising use case is as a supplement: a tool that can help children as part of a more holistic, teacher-led curriculum.

And the reasons for adding them to our tool kit are compelling: like animals, robots have some things to offer that human interaction doesn’t.

When a group of Yale researchers asked teachers who teach English to non-native speakers in K–2 classrooms to describe their biggest challenge, they said it was that the kids weren’t willing to make mistakes in front of the class. It was hindering their learning progress because the teachers couldn’t assess where the children needed help. Even in one-on-one interactions, the kids didn’t want to say the wrong answers. It was too embarrassing. The Yale researchers took a simple robot called Keepon that looked like a plump yellow cartoon bird and programmed it to interact with the kids and engage them in language conversations. Their hunch proved right: the kids worried less about making mistakes when talking to the robots. If we can add well-designed robots to an educational curriculum as a supplemental activity, it may increase children’s learning progress.

Even adults can feel nervous in other people’s presence and perform worse in experiments with games and math problems. We saw earlier that some people perform better with their dog in the room than alone or with a supportive friend. According to the early-stage research in HRI, robots are able to provide similar encouragement without judgment, even for adults. They can coach, motivate, and be a partner in sensitive or embarrassing tasks. It makes sense: we know that robots, like animals, can’t judge or evaluate us the way a person would. This lets robots take on certain support roles that humans can’t.

Nowhere is this supplemental effect more apparent than in the research on robot-assisted therapy for autism spectrum disorders (ASD). The first time I watched a video taken during this research, I was floored. An adolescent boy sat, slouching slightly, in a chair by a table in a sparse room. His body was turned away from the woman, his therapist, who was seated beside him. He looked away as she asked him questions, and he answered while avoiding eye contact. The boy, who struggles with recognition and expression in communication, was in a fairly standard therapy session with the woman he had worked with for many years. The only difference was that the researchers behind the one-way mirror in the room were about to bring in a robot.

When they placed the cat-sized baby dinosaur robot on the table in front of him, he was transfixed. He smiled as the robot shuffled toward him. And then, he glanced at his therapist ever so briefly, raising an eyebrow at her, as if to say, “Are you seeing this?” before turning his gaze back to the robot. This type of glance is called social referencing. It’s something this child rarely did in therapy or even at home with his own parents.

The boy watched as the robot dinosaur hesitated in front of a cartoon river drawn on the table. He’d been told that his job was to encourage the robot. “Ahhh,” said the boy. “You can do it!” He clenched his fists supportively, leaning toward the table. “Cross that river, come on!” he said enthusiastically. The robot walked forward. The boy glanced at the therapist again, shaking in silent mirth. She asked him, kindly: “Is it funny?” “Yeah,” he said, and he looked her straight in the eye as he continued, “especially when he got stuck.”

The robot encounter lasted only five minutes. The boy’s parents and the researchers behind the one-way mirror at the side of the room were thrilled to see the verbalizations and the eye contact, the “social referencing” that was so rare for this child. The researchers had been exploring the use of robots in therapy, hoping to see a positive response, and they were amazed at how much of one they were getting. Most importantly, the effect lasted beyond the interaction with the robot. By the time the robot was taken out of the room, the boy had turned to face his therapist and was talking to her. He was asking questions while looking right at her. And even though the robot was no longer there to “mediate” the interaction, he continued to communicate with his therapist in a way that the researchers and parents hadn’t seen in any of their previous sessions.

The modern version of Levinson and his therapy dog, Jingles, is Brian Scassellati, a Yale professor known affectionately in the robotics community as “Scaz.” The above study, designed by his student Eli Kim, used a robot in a role similar to Jingles, the animal therapist. Scaz’s research group, together with many other collaborators, has conducted hundreds of studies exploring the use of social robots in therapy and education. For nearly two decades, they’ve looked at using a variety of different robots in difficult situations for children, and with promising results. The robots consistently perform better than people or screens.

Excerpted from The New Breed: What Our History With Animals Reveals About Our Future With Robots by Kate Darling. Published by Henry Holt and Company. Copyright © 2021 by Kate Darling. All rights reserved.