Socially Assistive Robots Could Make You Healthier, Not Jobless

A growing field of research and development is designing intelligent, socially interactive machines that can help people with health education, physical rehabilitation, and elder care.

There's a lot of talk these days about the dangers of automation and the threat of robots stealing our jobs. But new research published today in the journal Science Robotics imagines a brighter future where robots train us, heal us, and help us get back to work.

Socially assistive robotics, or SAR, is a growing field of research and development in which engineers design intelligent, socially interactive machines that can help people in specific circumstances. As a field, it can be considered an offshoot of the more established field of medical rehabilitation robots.

Unlike traditional rehab machines that physically interact with patients - providing mechanical resistance during a muscle exercise, say - SAR systems are deigned to be principally social in nature. These are the bots that will sit with you during long weeks of physical rehab, offering coaching, monitoring, and companionship. SAR bots could also be deployed in elder care facilities or early childhood education.

In fact, several different kinds of robots have already been tested in these areas, said Maja Matarić, director of USC Robotics Research Lab and author of the Science paper. Her lab has participated in studies using SAR bots in stroke rehabilitation, mental health care, and autism. She's also used more kid-friendly SAR bots to help children learn about health and nutrition.

RELATED: This Rolling Robot Porter Follows You With Your Stuff

The advantages of SAR helper bots are readily apparent, Matarić said. Studies show that patients - and especially kids - tend to engage with robots better than traditional computer screens or tablets. The trick is developing robots that patients will feel comfortable, well, hanging with.

"Evidence from neuroscience shows that our brains respond with higher levels of activation to interactions with humans, pets, and robots than they do to screens," Matarić said. "So we are more engaged, learn more, and enjoy more interactions with those real, physical agents."

But first, engineers must get past the uncanny valley effect, that odd phenomenon where humans get creeped out by robots that are almost, but not quite, human. To address this dilemma, roboticists are working hard to improve the capacity of SAR bots to interact with humans. It's a unique design challenge involving physical appearance, speech recognition, body language, and the endless other elements that we unconsciously process during interaction and conversation.

"Evolutionary theories implicate complex social structures as a major driver of human intelligence," Matarić writes in the paper. "SAR designers must determine ways to achieve similar, compatible, socially interactive embodied systems that smoothly integrate the physical, cognitive, and social aspects of the robot."

The devil, as usual, is in the details. For example, research shows people get turned off if a robot responds too quickly in conversation - or too slowly. With humanoid bots, subtleties of facial expression, head orientation, eye contact, and verbal cadence all play a role. The challenges of human-machine interaction are one thing when dealing with recreational companion bots or digital personal assistants like Google Home or Amazon Echo.

But when you're designing hospital bots to help people though medical issues, the stakes are higher. If SAR bots are going to be genuinely helpful in therapeutic scenarios, Matarić said, the field will need input from dozens of disciplines in medicine, psychology, computer science, and engineering. She's particularly interested in getting young people into the field.

"Robotics is developing and changing and growing so quickly, that whatever one does in school today is going to be just the foundation for what will be possible when they get out to actually create real machines for people," she said.

WATCH: Why You Shouldn't Fear AI