global tech

Companion Robots Are Helping Autistic Children Feel Comfortable in School

Photo: BeatBots

In a special-needs primary school in Qatar’s capital, Doha, a 3-year-old girl on the autism spectrum has a new friend called Keepon. The bond has been five months in the making, since, like many children with autism spectrum disorder (ASD), Ghalia is wary of strangers. Keepon has been patient with Ghalia, weathering tantrums when they were first getting to know each other and not taking it personally when she wouldn’t make eye contact at first. Now, it’s clear Ghalia is fond of her new friend: She tops Keepon’s head with a knit cap and offers a kiss. The fact that Keepon doesn’t have a mouth or eyebrows doesn’t deter Ghalia. It’s actually by design. Keepon, yellow, snowman-shaped, and engineered by Hideki Kozima in Japan, is one of almost two dozen socially assistive robots deployed into Doha’s classrooms to help autistic children engage, interact, and learn. For the past five years, bio-roboticist Dr. John-John Cabibihan and colleagues at Qatar University have been at the forefront of developing culturally sensitive robotics for the Middle East’s growing autism population.

Cabibihan points to the rising rates of autism across the globe as “the driver” for his work. About 1.1 percent of children in Qatar are estimated to have ASD. Though there’s tremendous variability case to case, autistic children generally struggle with social cues and interpreting facial expressions. They may be resistant to holding eye contact and become easily frustrated. Their sensitivities can become disruptive to their education. Social robots, he believes — and research supports — can help.

The goal is to get the autistic child to engage, interact, imitate, and play. That might be as simple as meeting a robot’s eyes, or mirroring a movement the robot performs. “Since the robot is less complex in its appearance, nonverbal behavior, and movements than a human, the child is attracted to it. Then, that can be an entry point for the therapist to introduce the behavior she wants the child to learn,” says Cabibihan, who leads the university’s program in Doha. Autistic children respond better to technical toys than simple ones, and Cabibihan’s work takes those findings a step further: His robots act as tools or mediators between the teacher and the child, helping them meet in the middle and communicate in a language one another can understand.

Hifza Javed, an engineer who worked with Cabibihan on dispatching the robots to special-needs classrooms in Doha, says that their robots bypass a sort of neurological traffic jam. “As humans, when we communicate with each other, we tend to use multiple channels of communication simultaneously,” — raised eyebrows, gesturing hands, a vocal emphasis on a specific syllable meant to convey sarcasm. “This complex communication is often overstimulating for children. This is exactly where social robots can be instrumental: They are small-size and friendly, and they communicate simply with uncomplicated words and easy-to-interpret instructions.”

The robots, which are usually about the same size as the children, lack facial features like eyebrows and mobile cheekbones, which make their signaling and instructions more approachable and straightforward to those on the autism spectrum. “The children tend to form quick bonds with the robots and treat them as their social peers. This is the basis of social robots as intervention tools,” says Javed.

So what does this look like? Since many autistic children experience oral hypersensitivity, Cabibihan and Javed programmed Nao, a humanoid robot, to lift a toothbrush and brush its teeth, while delivering encouraging instructions for children to do the same. With Keepon, Ghalia mimicked the robot’s dancing movements, which required her to watch and follow instructions. A robot might be programmed to kick a ball, and a child must wait his turn among a group of peers to kick it back. All of these illustrate important behavioral and educational milestones for autistic children.

While Cabibihan’s work is not singular (similar research initiatives are underway in the U.K., at Yale, and in Luxembourg), there is a need for cultural sensitivity, even when it comes to robotics. “We are all getting similar results” regarding robotics benefits as a therapeutic intervention for ASD, he says. However, “language and different cultures” must be considered. Robots must be accessible to children who do not speak English but speak Arabic, for example.

Then, there’s what he calls the “form factor.” “For some who may not be familiar with the Middle East, they might introduce pigs or dogs — something that has religious or cultural implications for the region,” he says. “That can get you into hot water.” To address this issue, Javed programmed the tooth-brushing robot to perform a traditional Arab dance, clothed in a thob and headgear, while holding a sword. It went over so well that researchers want to use the robot in further imitation exercises.

Robotics as a meaningful intervention for autism has progressed within the last decade, but is not widely available. With the exception of Keepon, which retails for $100, the cost is exorbitant: The tooth-brushing, sword-dancing humanoid Nao costs $7,990 and furry seal Paro is $6,400. It’s something that Dr. Uvais Qidwai, a colleague of Cabibihan’s at Qatar University specializing in robotic teaching toys, is mindful of. “One of my students once commented on my work, ‘Does this mean that your solutions will be available only for the rich ones?’ and ever since, I have been trying to make robotic solutions that are lower in cost, so parents can afford them.” A bright spot is that since this technology is new, “we can make discoveries every day,” says Cabibihan.

He envisions a future where his assistive technology can help autistic people live independent lives by making everyday tasks more accessible, such as navigating public transportation and withdrawing money from ATMs. “We want children to get integrated into society.” However, the discoveries go both ways: Qidwai and Cabibihan express implicit support for a neurodiversity perspective, wherein autism is considered just one of the many normal ways human brains can be wired. “Having autism is not bad. We can help them introduce their world to others, and at the same time, they give insights as to what’s possible for innovation,” says Cabibihan.

Perhaps the most futuristic element of this research is the hope the researchers feel, and in turn, witness in the therapists, parents, and students. “In one session, there was one child, [whose] teachers were very skeptical that he would cooperate in the activity,” remembered Qidwai. The boy wasn’t listening and was walking around, flapping his hands in a self-stimulating behavior common among people on the spectrum. “As soon as the teacher turned the robot on and it made a yawning sound for waking up, the child’s eyes were glued to it.” The boy played with the robot for half an hour without assistance, hugged it, and greeted it in a “traditional style where two people touch their foreheads together as a sign of love and respect.”

Qidwai was watching with the boy’s teachers over closed circuit in a different room. “When I looked at them, they had tears in their eyes! They said that it usually takes them 20 minutes to get the student to his desk. We decided to leave that robot at the school.” Of his work, Cabibihan mused: “One thing I have observed, learning about robots, is that the more you try to learn about them, the more you learn about humans.”

An earlier version of this story incorrectly stated Hifza Javed’s degree status. At this time, she is pursuing her Ph.D., but has not yet received it.

Companion Robots Are Helping Autistic Children in Schools