Here and Nao: Better Support
Nao, an artificial intelligence robot, with Renáta Tichá.
A University of Minnesota team is testing an artificial intelligence robot to address the shortage of support workers and fight loneliness among older people and those with disabilities.
Faculty, staff, and a graduate student from the University’s Institute on Community Integration and its Minnesota Robotics Department recently completed a pilot study with 17 residents at Pillars of Prospect Park, an intergenerational senior living community, who interacted over several weeks with a programmable humanoid robot named Nao.
“A lot of times when something new is introduced, there’s excitement for a short time and then not so much, but all the way through the test, there were residents who were excited to have Nao back again,” said Erin Erdahl, director of programming and partnerships for Pillars. “Having staff and residents involved in the planning and development made it impactful for our community.”
Researchers have also worked with Udac Inc., a Duluth provider of services to people with disabilities, and are exploring grant funding to test Nao’s usefulness in the workplace.
“There is potential for Nao to take on a variety of roles, from reminding people to take medications to answering repeated questions without the judgment people with disabilities sometimes face,” said Karen Herman, Udac’s executive director.
The insight gained from the elderly and disability communities will be used to further test and refine Nao’s future interactions, said Renáta Tichá, who is leading the work for ICI.
“One eye-opening thing was how people responded to working with the robot differently,” she said. “Some were amazed by it and got quite attached, seeking us out every time we came. Then there was a neutral group, and some just found it odd and said they’d rather have a cat. This technology can support some people but not others. We need to develop an assessment to figure out who is best suited to what it can do.”
Another surprise was finding some people willing to share their emotions and information with the robot that they wouldn’t share with other people, Tichá said.
Nikos Papanikolopoulos, director of the University’s Minnesota Robotics Institute, was surprised at how deep the interactions were and credited Tichá and her team with working closely with the test sites to achieve them. That team includes Vassilios Morellas, Haoyi Shi, Maryam Mahmoudi, and Brian Abery.
Before the test, the team visited the Pillars community to talk with staff and residents about Nao and the study's design.
“The way we worked together was really positive,” Erdahl said. “We had a good sense of the focus of the study, and I loved that they wanted feedback from older adults rather than just assuming what older people would want. Any time we can offer and work with advancements and new technology in a pilot is really beneficial.”
Residents seemed to enjoy their interactions, she said.
“As people got to know Nao and interact with it, I think they started to see the benefits, like helping to remind them of appointments. I don’t think they were considering whether it could alleviate social isolation, but it kept people engaged, and perhaps the rest may happen indirectly.”
And that is Nao’s most significant test, Papanikolopoulos said.
“We are deeply indebted to Renáta and her team for helping us get all of this data. Now, we need these systems to show empathy. Nao is responding to that to some degree, and the hope is it will grow,” he said. “This is the biggest challenge: How can we make a mathematical model for empathy, and how can we manage the expectations that come with that? We still have a long way to go, but I see this helping to democratize support systems for marginalized people in our society.”
During the pilot study, Nao interacted with a resident at Pillars of Prospect Park, an intergenerational senior living community in Minneapolis.