There is a robot prototype that is being tested in hospitals in Texas. Its name is Moxi, it’s mostly clad in white, has digital expressions to take the place of a flesh-and-bone face, and most importantly, does mundane tasks that otherwise preoccupy human nurses. Now, those nurses increasingly have more time to spend with patients.
Moxi is not trying to fully replicate human nurses. That’s not the idea. Moxi exists at the intersection of machines, artificial intelligence, and making healthcare work better. The company that designed and is testing Moxi, Diligent Robotics of Austin, Texas, is led by CEO Andrea Thomaz, a social robotics expert who recognized early that AI could go a long way toward building robots able to address human needs.
“Socially guided machine learning was the root of my thesis at MIT,” says Thomaz, who earned her PhD from the Massachusetts Institute of Technology and enjoyed stints as a robotics professor at Georgia Tech and the University of Texas Austin. She says the challenge was that machine learning and the programming of computers are very different from the ways that humans learn from each other. But by dipping into developmental psychology and how children learn, Thomaz crafted a robot that can learn through experience.
This learning is not just in how it’s programmed. The people who work with Moxi can teach it new skills. “It’s like how a tennis coach would show a player how to swing a racquet,” says Thomaz.
We’re accustomed to seeing robots in manufacturing environments, but applying that idea to healthcare might seem a very different scenario. And it is. But surgeons are already using machines for head and neck and urologic surgery. Exoskeletons are helping paralyzed people walk. In development are pharmaceutical dispensers that work like ATM machines.
Moxi is in league with these innovations, but with a major difference. Thomaz and her colleagues—Vivian Chu, an expert in applying machine learning algorithms with multimodal data, and Agata Rozga, a psychologist who combines computation and observational methods in healthcare settings—have designed this robot to deal with the many variables and dynamics found on hospital patient floors. This includes the presence and movements of people, varying schedules, such as when bed linens are changed, or when a new patient is admitted. The primary tasks of the current Moxi model are to take those linens to the laundry area, to deliver supplies to newly admitted patient rooms, and to deliver lab samples.
Each is a task that consumes healthcare professionals’ time, underutilizing them relative to their education and licensing. Give those repetitive jobs to a robot and the nursing staff then can do what they do best: human caregiving.
Moxi is imbued with more than just a moving arm, a hand that grabs, and a rolling, mobile base. The digital face has eyes that blink and wink, a voice that says hello, and sensors that enable safe navigation. The head and eyes move and blink to proactively communicate intentions.
“Moxi can coexist with people and be part of their team,” says Thomaz.
Thomaz’s considerable résumé—she’s been featured in cover stories in both Popular Science and MIT Technology Review magazines—has attracted funding for Moxi’s development. The National Science Foundation contributed $725,000 in two grants in 2016 and 2017, while four separate venture capital funds added $2.1 million in seed funding in 2018.
There are other applications for robots like Moxi, and other robotic enterprises are working on similar products. It’s likely we’ll soon see and interact with robots in coffee shops, offices, warehouses, and in retail settings, where robots retrieving items from storage might dramatically reduce actual floor space dedicated to shopping. Already, virtual digital assistants in the home are used to conduct internet searches, play music, and order goods, making human-machine interaction commonplace. But healthcare has pressing challenges, which is what drew Diligent Robotics to the industry. Moxi’s goal is to make hospitals work more efficiently.
According to the Texas Organization of Nurse Executives, the state will face a shortage of seventy thousand full-time registered nurses by the year 2020. The Robert Wood Johnson Foundation says that the nursing workforce nationwide needs to increase, due to growth and replacements, by more than 1 million by 2024, particularly in the South and West, as the baby boomer generation continues to age.
“Healthcare has one of the biggest needs for robotics,” says Thomaz. “We knew that we could build this for the greatest societal impact.”
Thomaz and her team shadowed nurses in hospitals for more than 150 hours to study their workplace challenges. She learned which tasks were right for robots—and something more. “We were so inspired by the passion of healthcare professionals,” she says. “Their jobs are hard. I want to help them.”
As for other companies’ robots also entering the healthcare workforce, Thomaz is unfazed. “Competition is great,” she says. “The main difference is we have arms with grippers. And we’re not trying to reinvent the hospital. It’s just about changing the workflow, letting Moxi do the routine tasks while nurses work with patients.”